NeoAccess Reference Guide

VERSION 4.4.1     (for old versions: 4.3.1   4.2.0   3.9.2)

Note: this library may be used independently from the rest of the Brain Annex technology stack; you can simply install it from the standard python repository (use your IDE installer, or do a:    pip install neoaccess )



5 classes:

NeoAccessCore , NeoAccess   Source code

NodeSpecs , CypherMatch , CypherUtils (helper classes)   Source code


Background Information: Using Neo4j with Python : the Open-Source Library "NeoAccess"

Tutorial 1     Tutorial 2     Tutorial 3 (Pandas import)

Unit tests (with pytest)


CONTENTS:

Class NeoAccessCore

Class NeoAccess

Class NodeSpecs

Class CypherMatch

Class CypherUtils


Class NeoAccessCore

    IMPORTANT : for versions 4.4 of the Neo4j database

    A thin wrapper around the Neo4j python connectivity library "Neo4j Python Driver",
    which is documented at: https://neo4j.com/docs/api/python-driver/current/api.html

    This "CORE" library allows the execution of arbitrary Cypher (query language) commands,
    and helps manage the complex data structures that they return.
    It may be used independently,
    or as the foundation of the higher-level child class, "NeoAccess"

    SECTIONS IN THIS CLASS:
        * INIT (constructor) and DATABASE CONNECTION
        * RUNNING GENERIC CYPHER QUERIES
    
nameargumentsreturns
__init__self, host=os.getenv("NEO4J_HOST"), credentials=(os.getenv("NEO4J_USER"), os.getenv("NEO4J_PASSWORD")), apoc=False, debug=False, autoconnect=True
        If unable to create a Neo4j driver object, raise an Exception
        reminding the user to check whether the Neo4j database is running

        :param host:        URL to connect to database with.
                                EXAMPLES: bolt://123.456.0.29:7687  ,  bolt://your_domain.com:7687  ,  neo4j://localhost:7687
                                DEFAULT: read from NEO4J_HOST environmental variable
        :param credentials: Pair of strings (tuple or list) containing, respectively, the database username and password
                                DEFAULT: read from NEO4J_USER and NEO4J_PASSWORD environmental variables
        :param apoc:        Flag indicating whether apoc library is used on Neo4j database to connect to
                                Notes: APOC, if used, must also be enabled on the database.
                                The only method currently requiring APOC is export_dbase_json()
        :param debug:       Flag indicating whether a debug mode is to be used :
                                if True, all the Cypher queries, and some additional info, will get printed
        :param autoconnect  Flag indicating whether the class should establish connection to database at initialization
        
nameargumentsreturns
connectselfNone
        Attempt to establish a connection to the Neo4j database, using the credentials stored in the object.
        In the process, create and save a driver object.
        
nameargumentsreturns
test_dbase_connectionselfNone
        Attempt to perform a trivial Neo4j query, for the purpose of validating
        whether a connection to the database is possible.
        A failure at start time is typically indicative of invalid credentials

        :return:    None
        
nameargumentsreturns
versionselfstr
        Return the version of the Neo4j driver being used.  EXAMPLE: "4.4.11"

        :return:    A string with the version number
        
nameargumentsreturns
closeselfNone
        Terminate the database connection.
        Note: this method is automatically invoked
              after the last operation included in "with" statements

        :return:    None
        
nameargumentsreturns
block_query_executionselfNone
        To activate a special testing-only only.
        All the Cypher queries will get printed (just like done in debug mode),
        but no database operations will actually be performed.
        Caution: many functions will fail validations on the results
                 of the query that wasn't executed.  This option should probably
                 be combined with an Exception catch

        :return: None
        
nameargumentsreturns
unblock_query_executionselfNone
        Terminate the special "BLOCK QUERY EXECUTION" test mode.  Cypher queries will be normally executed.
        The state of the "debug" mode (only affecting extra diagnostic printing) won't be affected.

        :return:    None
        

RUN GENERIC QUERIES

nameargumentsreturns
queryself, q: str, data_binding=None, single_row=False, single_cell="", single_column=""
        Run a Cypher query.  Best suited for Cypher queries that return individual values,
        but may also be used with queries that return nodes or relationships or paths - or nothing.

        Execute the query and fetch the returned values as a list of dictionaries.
        In cases of no results, return an empty list.
        A new session to the database driver is started, and then immediately terminated after running the query.

        ALTERNATIVES:
            * if the Cypher query returns nodes, and one wants to extract the internal Neo4j ID's or labels
              (in addition to all the properties and their values) then use query_extended() instead.

            * in case of queries that alter the database (and may or may not return values),
              use update_query() instead, in order to retrieve information about the effects of the operation

        :param q:               A string with a Cypher query
        :param data_binding:    An optional Cypher dictionary
                                EXAMPLE, assuming that the cypher string contains the substrings "$node_id":
                                        {'node_id': 20}

        :param single_row:      (OPTIONAL) If True, return a dictionary containing just the first (0-th) result row, if present,
                                    or None in case of no results.
                                    Note that if the query returns multiple records, the picked one
                                    will be arbitrary, unless an ORDER BY is included in the query
                                    EXAMPLES of returned values:
                                        {"brand": "Toyota", "color": "White"}
                                        {'n': {}}

        :param single_cell:     (OPTIONAL) Meant for situations where only 1 node (record) is expected,
                                    and one wants only 1 specific field of that record.
                                    If a string is provided, return the value of the field by that name
                                    in the first (0-th) retrieved record.
                                    In case of no nodes found, or a node that lacks a key with this name, None will be returned.
                                    Note that if the query returns multiple records, the picked one
                                    will be arbitrary, unless an ORDER BY is included in the query

        :param single_column:   (OPTIONAL) Name of the column of interest.
                                If provided, assemble a list (possibly empty)
                                from all the values of that particular column all records.
                                Note: can also be used to extract data from a particular node, for queries that return whole nodes

        :return:        If any of single_row, single_cell or single_column are True, see info under their entries.
                        If those arguments are all False, it returns a (possibly empty) list of dictionaries.
                        Each dictionary in the list will depend on the nature of the Cypher query.
                        EXAMPLES:
                            Cypher returns nodes (after finding or creating them): RETURN n1, n2
                                    -> list item such as {'n1': {'gender': 'M', 'patient_id': 123}
                                                          'n2': {'gender': 'F', 'patient_id': 444}}
                            Cypher returns attribute values that get renamed: RETURN n.gender AS client_gender, n.pid AS client_id
                                    -> list items such as {'client_gender': 'M', 'client_id': 123}
                            Cypher returns attribute values without renaming them: RETURN n.gender, n.pid
                                    -> list items such as {'n.gender': 'M', 'n.pid': 123}
                            Cypher returns a single computed value
                                    -> a single list item such as {"count(n)": 100}
                            Cypher returns a single relationship, with or without attributes: MERGE (c)-[r:PAID_BY]->(p)
                                    -> a single list item such as [{ 'r': ({}, 'PAID_BY', {}) }]
                            Cypher returns a path:   MATCH p= .......   RETURN p
                                    -> list item such as {'p': [ {'name': 'Eve'}, 'LOVES', {'name': 'Adam'} ] }
                            Cypher creates nodes (without returning them)
                                    -> empty list
        
nameargumentsreturns
query_extendedself, q: str, data_binding = None, flatten = False, fields_to_exclude = None[dict]
        Extended version of query(), meant to extract additional info
        for queries that return so-called Graph Data Types,
        i.e. nodes, relationships or paths,
        such as:    "MATCH (n) RETURN n"
                    "MATCH (n1)-[r]->(n2) RETURN r"

        For example, useful in scenarios where nodes were returned,
        and their Neo4j internal IDs and/or labels are desired
        (in addition to all the properties and their values)

        Unless the flatten flag is True, individual records are kept as separate lists.
            For example, "MATCH (b:boat), (c:car) RETURN b, c"
            will return a structure such as [ [b1, c1] , [b2, c2] ]  if flatten is False,
            vs.  [b1, c1, b2, c2]  if  flatten is True.  (Note: each b1, c1, etc, is a dictionary.)

        TODO:  Scenario to test:
            if b1 == b2, would that still be [b1, c1, b1(b2), c2] or [b1, c1, c2] - i.e. would we remove the duplicates?
            Try running with flatten=True "MATCH (b:boat), (c:car) RETURN b, c" on data like "CREATE (b:boat), (c1:car1), (c2:car2)"

        :param q:               A Cypher query
        :param data_binding:    An optional Cypher dictionary
                                EXAMPLE, assuming that the cypher string contains the substring "$age":
                                        {'age': 20}
        :param flatten:         Flag indicating whether the Graph Data Types need to remain clustered by record,
                                    or all placed in a single flattened list
        :param fields_to_exclude:   Optional list of strings with name of fields
                                    (in the database or special ones added by this function)
                                    that wishes to drop.  No harm in listing fields that aren't present

        :return:        A (possibly empty) list of dictionaries, if flatten is True,
                        or a list of list, if flatten is False.
                        Each item in the lists is a dictionary, with details that will depend on which Graph Data Types
                                were returned in the Cypher query.
                                EXAMPLE of *individual items* - for a returned NODE
                                    {'gender': 'M', 'age': 20, 'internal_id': 123, 'neo4j_labels': ['patient']}
                                EXAMPLE of *individual items* - for a returned RELATIONSHIP
                                    {'price': 7500, 'internal_id': 2,
                                     'neo4j_start_node': ,
                                     'neo4j_end_node': ,
                                     'neo4j_type': 'bought_by'}]
        
nameargumentsreturns
update_queryself, q: str, data_binding=Nonedict
        Run a Cypher query and return statistics about its actions (such number of nodes created, etc.)
        Typical use is for queries that update the database.
        If the query returns any values, a list of them is also made available, as the value of the key 'returned_data'.

        Note: if the query creates nodes and one wishes to obtain their internal database ID's,
              one can include Cypher code such as "RETURN id(n) AS internal_id" (where n is the dummy name of the newly-created node)

        EXAMPLE:  result = update_query("CREATE(n :CITY {name: 'San Francisco'}) RETURN id(n) AS internal_id")

                  result will be {'nodes_created': 1, 'properties_set': 1, 'labels_added': 1,
                                  'returned_data': [{'internal_id': 123}]
                                 } , assuming 123 is the internal database ID of the newly-created node

        :param q:           Any Cypher query, but typically one that doesn't return anything
        :param data_binding: Data-binding dictionary for the Cypher query

        :return:            A dictionary of statistics (counters) about the query just run
                            EXAMPLES -
                                {}      The query had no effect
                                {'nodes_deleted': 3}    The query resulted in the deletion of 3 nodes
                                {'properties_set': 2}   The query had the effect of setting 2 properties
                                {'relationships_created': 1}    One new relationship got created
                                {'returned_data': [{'internal_id': 123}]}  'returned_data' contains the results of the query,
                                                                        if it returns anything, as a list of dictionaries
                                                                        - akin to the value returned by query()
                                {'returned_data': []}  Gets returned by SET QUERIES with no return statement
                            OTHER KEYS include:
                                nodes_created, nodes_deleted, relationships_created, relationships_deleted,
                                properties_set, labels_added, labels_removed,
                                indexes_added, indexes_removed, constraints_added, constraints_removed
                                More info:  https://neo4j.com/docs/api/python-driver/current/api.html#neo4j.SummaryCounters
        
nameargumentsreturns
debug_query_printself, q :str, data_binding=None, method=NoneNone
        Print out the given Cypher query
        (and, optionally, its data binding and/or the name of the calling method)

        :param q:               String with Cypher query
        :param data_binding:    OPTIONAL dictionary
        :param method:          OPTIONAL string with the name of the calling method
                                    EXAMPLE:  "foo"
        :return:                None
        



Class NeoAccess

    IMPORTANT : for versions 4.4 of the Neo4j database

    High-level class to interface with the Neo4j graph database from Python.

    Tested on version 4.4 of Neo4j Community version, but should work with other 4.x versions, too.
    NOT tested on any other major version of Neo4j; in particular, NOT tested with version 5

    This class is a layer above its parent class "NeoAccessCore",
        and it provides a higher-level functionality for common database operations,
        such as lookup, creation, deletion, modification, import, indices, etc.

    SECTIONS IN THIS CLASS:
        * INTERNAL DATABASE ID
        * RETRIEVE DATA
        * FOLLOW LINKS
        * CREATE NODES
        * DELETE NODES
        * MODIFY FIELDS
        * RELATIONSHIPS
        * LABELS
        * INDEXES
        * CONSTRAINTS
        * READ IN DATA from PANDAS
        * JSON IMPORT/EXPORT
        * DEBUGGING SUPPORT

    It makes use of separate classes (NOT meant for the end user) in the file cypher_utils.py
    
nameargumentsreturns
assert_valid_internal_idself, internal_id: intNone
        Raise an Exception if the argument is not a valid database internal ID

        :param internal_id: Alleged Neo4j internal database ID
        :return:            None
        

RETRIEVE DATA

nameargumentsreturns
get_record_by_primary_keyself, labels: str, primary_key_name: str, primary_key_value, return_internal_id=FalseUnion[dict, None]
        Return the first (and it ought to be only one) record with the given primary key, and the optional label(s),
        as a dictionary of all its attributes.

        If more than one record is found, an Exception is raised.
        If no record is found, return None.

        :param labels:              A string or list/tuple of strings.  Use None if not to be included in search
        :param primary_key_name:    The name of the primary key by which to look the record up
        :param primary_key_value:   The desired value of the primary key
        :param return_internal_id:  If True, an extra entry is present in the dictionary, with the key "internal_id"

        :return:                    A dictionary, if a unique record was found; or None if not found
        
nameargumentsreturns
exists_by_keyself, labels: str, key_name: str, key_valuebool
        Return True if a node with the given labels and key_name/key_value exists, or False otherwise
        TODO: test for multiple labels
        :param labels:      A string or list/tuple of strings
        :param key_name:    A string with the name of a node attribute
        :param key_value:   The desired value of the key_name attribute
        :return:            True if a node with the given labels and key_name/key_value exists,
                                or False otherwise
        
nameargumentsreturns
exists_by_internal_idself, internal_idbool
        Return True if a node with the given internal database ID exists, or False otherwise

        :param internal_id: An integer with a node's internal database ID
        :return:            True if a node with the given internal Neo4j exists, or False otherwise
        
nameargumentsreturns
count_nodesselfint
        Compute and return the total number of nodes in the database
        TODO: test

        :return:    The total number of nodes in the database
        
nameargumentsreturns
get_single_fieldself, match: Union[int, NodeSpecs], field_name: str, order_by=None, limit=Nonelist
        For situations where one is fetching just 1 field,
        and one desires a list of the values of that field, rather than a dictionary of records.
        In other respects, similar to the more general get_nodes()

        :param match:       EITHER an integer with an internal database node id,
                                OR a "NodeSpecs" object, as returned by match(), with data to identify a node or set of nodes
        :param field_name:  A string with the name of the desired field (attribute)
        :param order_by:    see get_nodes()
        :param limit:       see get_nodes()

        :return:  A list of the values of the field_name attribute in the nodes that match the specified conditions
        
nameargumentsreturns
get_nodesself, match: Union[int, NodeSpecs], return_internal_id=False, return_labels=False, order_by=None, limit=None, single_row=False, single_cell=""
        RETURN a list of the records (as dictionaries of ALL the key/value node properties)
        corresponding to all the Neo4j nodes specified by the given match data.

        :param match:           EITHER an integer with an internal database node id,
                                    OR a "NodeSpecs" object, as returned by match(), with data to identify a node or set of nodes

        :param return_internal_id:  Flag indicating whether to also include the Neo4j internal node ID in the returned data
                                    (using "internal_id" as its key in the returned dictionary)
        :param return_labels:   Flag indicating whether to also include the Neo4j label names in the returned data
                                    (using "neo4j_labels" as its key in the returned dictionary)

        :param order_by:        (OPTIONAL) String with the key (field) name to order by, in ascending order
                                    Caution: lower and uppercase names are treated differently in the sort order
        :param limit:           (OPTIONAL) Integer to specify the maximum number of nodes returned

        :param single_row:      Meant in situations where only 1 node (record) is expected, or perhaps one wants to sample the 1st one;
                                    if not found, None will be returned [to distinguish it from a found record with no fields!]

        :param single_cell:     Meant in situations where only 1 node (record) is expected, and one wants only 1 specific field of that record.
                                If single_cell is specified, return the value of the field by that name in the first node
                                Note: this will be None if there are no results, or if the first (0-th) result row lacks a key with this name
                                TODO: test and give examples.  single_cell="name" will return result[0].get("name")

        :return:                If single_cell is specified, return the value of the field by that name in the first node.
                                If single_row is True, return a dictionary with the information of the first record (or None if no record exists)
                                Otherwise, return a (possibly-empty) list whose entries are dictionaries with each record's information
                                    (the node's attribute names are the keys)
                                    EXAMPLE: [  {"gender": "M", "age": 42, "condition_id": 3},
                                                {"gender": "M", "age": 76, "location": "Berkeley"}
                                             ]
                                    Note that ALL the attributes of each node are returned - and that they may vary across records.
                                    If the flag return_nodeid is set to True, then an extra key/value pair is included in the dictionaries,
                                            of the form     "internal_id": some integer with the Neo4j internal node ID
                                    If the flag return_labels is set to True, then an extra key/value pair is included in the dictionaries,
                                            of the form     "neo4j_labels": [list of Neo4j label(s) attached to that node]
                                    EXAMPLE using both of the above flags:
                                        [  {"internal_id": 145, "neo4j_labels": ["person", "client"], "gender": "M", "condition_id": 3},
                                           {"internal_id": 222, "neo4j_labels": ["person"], "gender": "M", "location": "Berkeley"}
                                        ]
        # TODO: provide an option to specify the desired fields

        
nameargumentsreturns
get_dfself, match: Union[int, NodeSpecs], order_by=None, limit=Nonepd.DataFrame
        Similar to get_nodes(), but with fewer arguments - and the result is returned as a Pandas dataframe

        [See get_nodes() for more information about the arguments]

        :param match:       EITHER an integer with an internal database node id,
                                OR a "NodeSpecs" object, as returned by match(), with data to identify a node or set of nodes
        :param order_by:    Optional string with the key (field) name to order by, in ascending order
                                Note: lower and uppercase names are treated differently in the sort order
        :param limit:       Optional integer to specify the maximum number of nodes returned

        :return:            A Pandas dataframe
        
nameargumentsreturns
matchself, labels=None, internal_id=None, key_name=None, key_value=None, properties=None, clause=None, dummy_node_name="n"NodeSpecs
        Return a "NodeSpecs" object storing all the passed specifications (the "RAW match structure"),
        as expected as argument in various other functions in this library,
        in order to identify a node or group of nodes.

        IMPORTANT:  if internal_id is provided, all other conditions are DISREGARDED;
                    otherwise, an implicit AND applies to all the specified conditions.

        Note:   NO database operation is actually performed by this function.

        [Other names explored: identify(), preserve(), define_match(), locate(), choose() or identify()]

        ALL THE ARGUMENTS ARE OPTIONAL (no arguments at all means "match everything in the database")
        :param labels:      A string (or list/tuple of strings) specifying one or more Neo4j labels.
                                (Note: blank spaces ARE allowed in the strings)
                                EXAMPLES:  "cars"
                                            ("cars", "powered vehicles")
                            Note that if multiple labels are given, then only nodes with ALL of them will be matched;
                            at present, there's no way to request an "OR" operation

        :param internal_id: An integer with the node's internal database ID.
                                If specified, it OVER-RIDES all the remaining arguments [except for the labels (TODO: revisit this)]

        :param key_name:    A string with the name of a node attribute; if provided, key_value must be present, too
        :param key_value:   The required value for the above key; if provided, key_name must be present, too
                                Note: no requirement for the key to be primary

        :param properties:  A (possibly-empty) dictionary of property key/values pairs, indicating a condition to match.
                                EXAMPLE: {"gender": "F", "age": 22}

        :param clause:      Either None, OR a (possibly empty) string containing a Cypher subquery,
                            OR a pair/list (string, dict) containing a Cypher subquery and the data-binding dictionary for it.
                            The Cypher subquery should refer to the node using the assigned dummy_node_name (by default, "n")
                                IMPORTANT:  in the dictionary, don't use keys of the form "n_par_i",
                                            where n is the dummy node name and i is an integer,
                                            or an Exception will be raised - those names are for internal use only
                                EXAMPLES:   "n.age < 25 AND n.income > 100000"
                                            ("n.weight < $max_weight", {"max_weight": 100})

        :param dummy_node_name: A string with a name by which to refer to the node (by default, "n");
                                only used if a `clause` argument is passed

        :return:            A python data dictionary, to preserve together all the passed arguments
        
nameargumentsreturns
get_node_internal_idself, match: NodeSpecsint
        Return the internal database ID of a SINGLE node identified by the "match" data
        created by a call to match().

        If not found, or if more than 1 found, an Exception is raised

        :param match:   A "NodeSpecs" object, as returned by match(), with data to identify a node or set of nodes
        :return:        An integer with the internal database ID of the located node,
                        if exactly 1 node is found; otherwise, raise an Exception
        
nameargumentsreturns
get_node_labelsself, internal_id: int[str]
        Return a list whose elements are the label(s) of the node specified by its Neo4j internal ID

        TODO: maybe also accept a "match" structure as argument

        :param internal_id: An integer with a Neo4j node id
        :return:            A list of strings with the names of all the labels of the given node
        

CREATE NODES

nameargumentsreturns
create_nodeself, labels, properties=Noneint
        Create a new node with the given label(s),
        and with the attributes/values specified in the properties dictionary.
        Return the Neo4j internal ID of the node just created.

        :param labels:      A string, or list/tuple of strings, specifying Neo4j labels (ok to have blank spaces);
                                it's acceptable to be None
        :param properties:  OPTIONAL (possibly empty or None) dictionary of properties to set for the new node.
                                EXAMPLE: {'age': 22, 'gender': 'F'}

        :return:            An integer with the internal database ID of the node just created
        
nameargumentsreturns
merge_nodeself, labels, properties=Nonedict
        The node gets created only if no other node with same labels and properties exists.

        Create a new node with the given label(s) and with the attributes/values specified in the properties dictionary.

        :param labels:      A string, or list/tuple of strings, specifying Neo4j labels (ok to have blank spaces)
        :param properties:  An optional (possibly empty or None) dictionary of properties
                                to try to match in an existing node, or - if not found - to set in a new node.
                                EXAMPLE: {'age': 22, 'gender': 'F'}

        :return:            A dict with 2 keys: "created" (True if a new node was created, or False otherwise)
                                                and "internal_id"
        
nameargumentsreturns
create_attached_nodeself, labels, properties = None, attached_to = None, rel_name = None, rel_dir = "OUT", merge=Trueint
        Create a new node (or possibly re-use an existing one),
        with the given labels and optional specified properties,
        and attached all the EXISTING nodes specified in the (possibly empty) list of nodes attached_to,
        using the given relationship name.
        All the relationships are OUTbound or INbound from the newly-created node,
        depending on the value of rel_dir.

        If merge=True, if an existing node is already present with the same labels and properties,
        it will be re-used rather than created (in that case, only the relationships will be created)

        Note: this is a simpler version of create_node_with_links()

        If any of the requested link nodes isn't found,
        then no new node is created, and an Exception is raised.

        Note: under unusual circumstances, the new node may be created even in situations where Exceptions are raised;
              for example, if attempting to create two identical relationships to the same existing node.

        EXAMPLE:
            create_attached_node(
                                    labels="COMPANY",
                                    properties={"name": "Acme Gadgets", "city": "Berkeley"},
                                    attached_to=[123, 456],
                                    rel_name="EMPLOYS"
            )

        :param labels:      Labels to assign to the newly-created node (a string, possibly empty, or list of strings)
        :param properties:  (OPTIONAL) A dictionary of optional properties to assign to the newly-created node
        :param attached_to: (OPTIONAL) An integer, or list/tuple of integers,
                                with internal database ID's to identify the existing nodes;
                                use None, or an empty list, to indicate if there aren't any
        :param rel_name:    (OPTIONAL) Name of the newly created relationships.
                                This is required, if an attached_to list was provided
        :param rel_dir:     (OPTIONAL) Either "OUT"(default), "IN" or "BOTH".  Direction(s) of the relationships to create
        :param merge:       (OPTIONAL) If True (default), a new node gets created only if there's no existing node
                                with the same properties and labels

        :return:            An integer with the internal database ID of the newly-created node
        
nameargumentsreturns
create_node_with_linksself, labels, properties=None, links=None, merge=Falseint
        Create a new node, with the given labels and optional properties,
        and link it up to all the EXISTING nodes that are specified
        in the (possibly empty) list of link nodes, identified by their Neo4j internal ID's.

        The list of link nodes also contains the names to give to each link,
        as well as their directions (by default OUTbound from the newly-created node)
        and, optionally, properties on the links.

        If any of the requested link nodes isn't found,
        then no new node is created, and an Exception is raised.

        Note: the new node may be created even in situations where Exceptions are raised;
              for example, if attempting to create two identical relationships to the same existing node.

        EXAMPLE (assuming the nodes with the specified Neo4j IDs already exist):
            create_node_with_links(
                                labels="PERSON",
                                properties={"name": "Julian", "city": "Berkeley"},
                                links=[ {"internal_id": 123, "rel_name": "LIVES IN"},
                                        {"internal_id": 456, "rel_name": "EMPLOYS", "rel_dir": "IN"},
                                        {"internal_id": 789, "rel_name": "OWNS", "rel_attrs": {"since": 2022}}
                                      ]
            )

        :param labels:      Labels to assign to the newly-created node (optional but recommended):
                                a string or list/tuple of strings; blanks allowed inside strings
        :param properties:  A dictionary of optional properties to assign to the newly-created node
        :param links:       Optional list of dicts identifying existing nodes,
                                and specifying the name, direction and optional properties
                                to give to the links connecting to them;
                                use None, or an empty list, to indicate if there aren't any
                                Each dict contains the following keys:
                                    "internal_id"   REQUIRED - to identify an existing node
                                    "rel_name"      REQUIRED - the name to give to the link
                                    "rel_dir"       OPTIONAL (default "OUT") - either "IN" or "OUT" from the new node
                                    "rel_attrs"     OPTIONAL - A dictionary of relationship attributes
        :param merge:       (OPTIONAL; default False) If True, a new node gets created only if there's no existing node
                                with the same properties and labels     TODO: test more

        :return:            An integer with the Neo4j ID of the newly-created node
        
nameargumentsreturns
_assemble_query_for_linkingself, links: listtuple
        Helper function for create_node_with_links(), and perhaps future methods.

        Given a list of existing nodes, and info on links to create to/from them,
        define the portions of the Cypher query to locate the existing nodes,
        and to link up to them.
        No query is actually run.

        :param links:   A list: SEE explanation in create_node_with_links()
        :return:        A 4-tuple with the parts of the query, as well as the needed data binding
                            1) q_MATCH
                            2) q_WHERE
                            3) q_MERGE
                            4) data_binding
        
nameargumentsreturns
create_node_with_relationshipsself, labels, properties=None, connections=Noneint
        TODO: this method may no longer be needed, given the new method create_node_with_links()
              Maybe ditch, or extract the Neo4j ID's from the connections,
              and call create_node_with_links()

        Create a new node with relationships to zero or more PRE-EXISTING nodes
        (identified by their labels and key/value pairs).

        If the specified pre-existing nodes aren't found, then no new node is created,
        and an Exception is raised.

        On success, return the Neo4j internal ID of the new node just created.

        Note: if all connections are in one direction, and with same (property-less) relationship name,
              and to nodes with known Neo4j internal IDs, then
              the simpler method create_attached_node() may be used instead

        EXAMPLE:
            create_node_with_relationships(
                                            labels="PERSON",
                                            properties={"name": "Julian", "city": "Berkeley"},
                                            connections=[
                                                        {"labels": "DEPARTMENT",
                                                         "key": "dept_name", "value": "IT",
                                                         "rel_name": "EMPLOYS", "rel_dir": "IN"},

                                                        {"labels": ["CAR", "INVENTORY"],
                                                         "key": "vehicle_id", "value": 12345,
                                                         "rel_name": "OWNS", "rel_attrs": {"since": 2021} }
                                            ]
            )

        :param labels:      A string, or list of strings, with label(s) to assign to the new node
        :param properties:  A dictionary of properties to assign to the new node
        :param connections: A (possibly empty) list of dictionaries with the following keys
                            (all optional unless otherwise specified):
                                --- Keys to locate an existing node ---
                                    "labels"        RECOMMENDED
                                    "key"           REQUIRED
                                    "value"         REQUIRED
                                --- Keys to define a relationship to it ---
                                    "rel_name"      REQUIRED.  The name to give to the new relationship
                                    "rel_dir"       Either "OUT" or "IN", relative to the new node (by default, "OUT")
                                    "rel_attrs"     A dictionary of relationship attributes

        :return:            If successful, an integer with the Neo4j internal ID of the node just created;
                                otherwise, an Exception is raised
        

DELETE NODES

nameargumentsreturns
delete_nodesself, match: Union[int, NodeSpecs]int
        Delete the node or nodes specified by the match argument.
        Return the number of nodes deleted.

        :param match:   EITHER an integer with an internal database node id,
                            OR a "NodeSpecs" object, as returned by match(), with data to identify a node or set of nodes
        :return:        The number of nodes deleted (possibly zero)
        
nameargumentsreturns
delete_nodes_by_labelself, delete_labels=None, keep_labels=NoneNone
        Empty out (by default completely) the Neo4j database.
        Optionally, only delete nodes with the specified labels, or only keep nodes with the given labels.
        Note: the keep_labels list has higher priority; if a label occurs in both lists, it will be kept.
        IMPORTANT: it does NOT clear indexes; "ghost" labels may remain!
        TODO: return the number of nodes deleted

        :param delete_labels:   An optional string, or list of strings, indicating specific labels to DELETE
        :param keep_labels:     An optional string or list of strings, indicating specific labels to KEEP
                                    (keep_labels has higher priority over delete_labels)
        :return:                None
        
nameargumentsreturns
bulk_delete_by_labelself, label: str
        IMPORTANT: APOC required (starting from v 4.4 of Neo4j, will be able to do this without APOC; TODO: not yet tested)

        Meant for large databases, where the straightforward deletion operations may result
        in very large number of nodes, and take a long time (or possibly fail)

        "If you need to delete some large number of objects from the graph,
        one needs to be mindful of the not building up such a large single transaction
        such that a Java OUT OF HEAP Error will be encountered."
        See:  https://neo4j.com/developer/kb/large-delete-transaction-best-practices-in-neo4j/

        TODO: generalize to bulk-deletion not just by label
        TODO: test.  CAUTION: only tested interactively

        :param label:   A string with the label of the nodes to delete (blank spaces in name are ok)
        :return:        A dict with the keys "batches" and "total"
        
nameargumentsreturns
empty_dbaseself, keep_labels=None, drop_indexes=False, drop_constraints=FalseNone
        Use this to get rid of everything in the database,
        including all the indexes and constraints (unless otherwise specified.)
        Optionally, keep nodes with a given label, or keep the indexes, or keep the constraints

        :param keep_labels:     An optional list of strings, indicating specific labels to KEEP
        :param drop_indexes:    Flag indicating whether to also ditch all indexes (by default, True)
        :param drop_constraints:Flag indicating whether to also ditch all constraints (by default, True)

        :return:                None
        

MODIFY FIELDS

nameargumentsreturns
set_fieldsself, match: Union[int, NodeSpecs], set_dict: dict int
        EXAMPLE - locate the "car" with vehicle id 123 and set its color to "white" and price to 7000
            match_structure = match(labels = "car", properties = {"vehicle id": 123})
            set_fields(match=match_structure, set_dict = {"color": "white", "price": 7000})

        NOTE: other fields are left un-disturbed

        Return the number of properties set.

        :param match:       EITHER an integer with an internal database node id,
                                OR a "NodeSpecs" object, as returned by match(), with data to identify a node or set of nodes
        :param set_dict:    A dictionary of field name/values to create/update the node's attributes
                            (note: blanks ARE allowed in the keys)

        :return:            The number of properties set
        

RELATIONSHIPS

nameargumentsreturns
get_relationship_typesself[str]
        Extract and return a list of all the Neo4j relationship names (i.e. types of relationships)
        present in the entire database, in no particular order

        :return:    A list of strings
        
nameargumentsreturns
add_linksself, match_from: Union[int, NodeSpecs], match_to: Union[int, NodeSpecs], rel_name:strint
        Add one or more links (aka graph edges/relationships), with the specified rel_name,
        originating in each of the nodes specified by the match_from specifications,
        and terminating in each of the nodes specified by the match_to specifications

        Return the number of links added; if none were added, or in case of error, raise an Exception.

        Notes:  - if a relationship with the same name already exists, nothing gets created (and an Exception is raised)
                - more than 1 node could be present in either of the matches

        TODO: add a `rel_props` argument
              (Unclear what multiple calls would do in this case: update the props or create a new relationship??)

        :param match_from:  EITHER an integer with an internal database node id,
                                OR a "NodeSpecs" object, as returned by match(), with data to identify a node or set of nodes
        :param match_to:    EITHER an integer with an internal database node id,
                                OR a "NodeSpecs" object, as returned by match(), with data to identify a node or set of nodes
                            Note: match_from and match_to, if created by calls to match(),
                                  in scenarios where a clause dummy name is actually used,
                                  MUST use different clause dummy names.

        :param rel_name:    The name to give to all the new relationships between the 2 specified nodes, or sets or nodes.
                                Blanks allowed.

        :return:            The number of edges added.  If none got added, or in case of error, an Exception is raised
        
nameargumentsreturns
add_links_fastself, match_from: int, match_to: int, rel_name:strint
        Method optimized for speed.  Only internal database ID are used.

        Add a links (aka graph edges/relationships), with the specified rel_name,
        originating in the node identified by match_from,
        and terminating in the node identified by match_to

        :param match_from:  An integer with an internal Neo4j node id
        :param match_to:    An integer with an internal Neo4j node id
        :param rel_name:    The name to give to the new relationship between the 2 specified nodes.  Blanks allowed

        :return:            The number of links added.  If none got added, or in case of error, an Exception is raised
        
nameargumentsreturns
remove_linksself, match_from: Union[int, NodeSpecs], match_to: Union[int, NodeSpecs], rel_nameint
        Remove one or more links (aka relationships/edges)
        originating in any of the nodes specified by the match_from specifications,
        and terminating in any of the nodes specified by the match_to specifications,
        optionally matching the given relationship name (will remove all edges if the name is blank or None)

        Return the number of edges removed; if none found, or in case of error, raise an Exception.

        Notes: - the nodes themselves are left untouched
               - more than 1 node could be present in either of the matches
               - the number of relationships deleted could be more than 1 even with a single "from" node and a single "to" node;
                        Neo4j allows multiple relationships with the same name between the same two nodes,
                        as long as the relationships differ in their properties

        :param match_from:  EITHER an integer with an internal database node id,
                                OR a "NodeSpecs" object, as returned by match(), with data to identify a node or set of nodes
        :param match_to:    EITHER an integer with an internal database node id,
                                OR a "NodeSpecs" object, as returned by match(), with data to identify a node or set of nodes
                            Note: match_from and match_to, if created by calls to match(),
                                  in scenarios where a clause dummy name is actually used,
                                  MUST use different clause dummy names.

        :param rel_name:    (OPTIONAL) The name of the relationship to delete between the 2 specified nodes;
                                if None or a blank string, all relationships between those 2 nodes will get deleted.
                                Blanks allowed.

        :return:            The number of edges removed.  If none got deleted, or in case of error, an Exception is raised
        
nameargumentsreturns
links_existself, match_from: Union[int, NodeSpecs], match_to: Union[int, NodeSpecs], rel_name: strbool
        Return True if one or more edges (relationships) with the specified name exist in the direction
        from and to the nodes (individual nodes or set of nodes) specified in the first two arguments.
        Typically used to find whether 2 given nodes have a direct link between them.

        :param match_from:  EITHER an integer with an internal database node id,
                                OR a "NodeSpecs" object, as returned by match(), with data to identify a node or set of nodes
        :param match_to:    EITHER an integer with an internal database node id,
                                OR a "NodeSpecs" object, as returned by match(), with data to identify a node or set of nodes
                            Note: match_from and match_to, if created by calls to match(),
                                  in scenarios where a clause dummy name is actually used,
                                  MUST use different clause dummy names.

        :param rel_name:    The name of the relationship to look for between the 2 specified nodes.
                                Blanks are allowed

        :return:            True if one or more relationships were found, or False if not
        
nameargumentsreturns
number_of_linksself, match_from: Union[int, NodeSpecs], match_to: Union[int, NodeSpecs], rel_name: strint
        Return the number of links (aka edges or relationships) with the given name
        that exist in the direction from and to the nodes (individual nodes or set of nodes)
        that are specified in the first two arguments.

        :param match_from:  EITHER an integer with an internal database node id,
                                OR a "NodeSpecs" object, as returned by match(), with data to identify a node or set of nodes
        :param match_to:    EITHER an integer with an internal database node id,
                                OR a "NodeSpecs" object, as returned by match(), with data to identify a node or set of nodes
                            Note: match_from and match_to, if created by calls to match(),
                                  in scenarios where a clause dummy name is actually used,
                                  MUST use different clause dummy names.

        :param rel_name:    The name of the relationship to look for between the 2 specified nodes or groups of nodes.
                                Blanks are allowed

        :return:            The number of links (relationships) that were found
        
nameargumentsreturns
reattach_nodeself, node, old_attachment, new_attachment, rel_name:str, rel_name_new=NoneNone
        Sever the relationship with the given name from the given node to the node old_attachment,
        and re-create it to the node new_attachment (optionally under a different relationship name).

        Note: relationship properties, if present, will NOT be transferred

        :param node:            An integer with the internal database ID of the node to detach and reattach
        :param old_attachment:  An integer with the internal database ID of the other node currently connected to
        :param new_attachment:  An integer with the internal database ID of the new node to connect to
        :param rel_name:        Name of the old relationship name
        :param rel_name_new:    (OPTIONAL) Name of the new relationship name (by default the same as the old one)

        :return:                None.  If unsuccessful, an Exception is raised
        
nameargumentsreturns
link_nodes_by_idsself, node_id1:int, node_id2:int, rel:str, rel_props = NoneNone
        Locate the pair of Neo4j nodes with the given Neo4j internal ID's.
        If they are found, add a relationship - with the name specified in the rel argument,
        and with the specified optional properties - from the 1st to 2nd node - unless already present.

        EXAMPLE:    link_nodes_by_ids(123, 88, "AVAILABLE_FROM", {'cost': 1000})

        TODO: maybe return a status, or the Neo4j ID of the relationship just created

        :param node_id1:    An integer with the Neo4j internal ID to locate the 1st node
        :param node_id2:    An integer with the Neo4j internal ID to locate the 2nd node
        :param rel:         A string specifying a Neo4j relationship name
        :param rel_props:   Optional dictionary with the relationship properties.  EXAMPLE: {'since': 2003, 'code': 'xyz'}
        :return:            None
        
nameargumentsreturns
link_nodes_on_matching_propertyself, label1:str, label2:str, property1:str, rel:str, property2=NoneNone
        Locate any pair of Neo4j nodes where all of the following hold:
                            1) the first one has label1
                            2) the second one has label2
                            3) the two nodes agree in the value of property1 (if property2 is None),
                                        or in the values of property1 in the 1st node and property2 in the 2nd node
        For any such pair found, add a relationship - with the name specified in the rel argument - from the 1st to 2nd node,
        unless already present.

        This operation is akin to a "JOIN" in a relational database; in pseudo-code:
                "WHERE label1.value(property1) = label2.value(property1)"       # if property2 is None
                    or
                "WHERE label1.value(property1) = label2.value(property2)"

        :param label1:      A string against which the label of the 1st node must match
        :param label2:      A string against which the label of the 2nd node must match
        :param property1:   Name of property that must be present in the 1st node (and also in 2nd node, if property2 is None)
        :param property2:   Name of property that must be present in the 2nd node (may be None)
        :param rel:         Name to give to all relationships that get created
        :return:            None
        

FOLLOW LINKS

nameargumentsreturns
follow_linksself, match: Union[int, NodeSpecs], rel_name: str, rel_dir ="OUT", neighbor_labels = None[dict]
        From the given starting node(s), follow all the relationships to or from it,
        from/into neighbor nodes (optionally having the given labels),
        and return all the properties of those neighbor nodes.

        :param match:           EITHER an integer with an internal database node id,
                                    OR a "NodeSpecs" object, as returned by match(), with data to identify a node or set of nodes
        :param rel_name:        A string with the name of relationship to follow.  (Note: any other relationships are ignored)
        :param rel_dir:         Either "OUT"(default), "IN" or "BOTH".  Direction(s) of the relationship to follow
        :param neighbor_labels: Optional label(s) required on the neighbors.
                                    If used, pass either a string or list of strings

        :return:                A list of dictionaries with all the properties of the neighbor nodes
                                TODO: maybe add the option to just return a subset of fields
        
nameargumentsreturns
count_linksself, match: Union[int, NodeSpecs], rel_name: str, rel_dir="OUT", neighbor_labels = Noneint
        From the given starting node(s), count all the relationships OF THE GIVEN NAME to and/or from it,
        into/from neighbor nodes (optionally having the given labels)

        :param match:           EITHER an integer with an internal database node id,
                                    OR a "NodeSpecs" object, as returned by match(), with data to identify a node or set of nodes
        :param rel_name:        A string with the name of relationship to follow.  (Note: any other relationships are ignored)
        :param rel_dir:         Either "OUT"(default), "IN" or "BOTH".  Direction(s) of the relationship to follow
        :param neighbor_labels: Optional label(s) required on the neighbors.  If present, either a string or list of strings

        :return:                The total number of inbound and/or outbound relationships to the given node(s)
        
nameargumentsreturns
get_parents_and_childrenself, internal_id: int()
        Fetch all the nodes connected to the given one by INbound relationships to it (its "parents"),
        as well as by OUTbound relationships to it (its "children")
        TODO: allow specifying a relationship name to follow

        :param internal_id: An integer with a Neo4j internal node ID
        :return:            A dictionary with 2 keys: 'parent_list' and 'child_list'
                                The values are lists of dictionaries with 3 keys: "internal_id", "label", "rel"
                                EXAMPLE of individual items in either parent_list or child_list:
                                {'internal_id': 163, 'labels': ['Subject'], 'rel': 'HAS_TREATMENT'}
        
nameargumentsreturns
get_siblingsself, internal_id: int, rel_name: str, rel_dir="OUT", order_by=None[int]
        Return the data of all the "sibling" nodes of the given one.
        By "sibling", we mean: "sharing a link (by default outbound) of the specified name,
        to a common other node".

        EXAMPLE: 2 nodes, "French" and "German",
                 each with a outbound link named "subcategory_of" to a third node,
                 will be considered "siblings" under rel_name="subcategory_of" and rel_dir="OUT

        :param internal_id: Integer with the internal database ID of the node of interest
        :param rel_name:    The name of the relationship used to establish a "siblings" connection
        :param rel_dir:     (OPTIONAL) Either "OUT" (default) or "IN".  The link direction that is expected from the
                                start node to its "parents" - and then IN REVERSE to the parent's children
        :param order_by:    (OPTIONAL) If specified, it must be the name of a field in
                                the sibling nodes, to order the results by; capitalization is ignored
        :return:            A list of dictionaries, with one element for each "sibling";
                                each element contains the 'internal_id' and 'neo4j_labels' keys,
                                plus whatever attributes are stored on that node.
                                EXAMPLE of single element:
                                {'name': 'French', 'internal_id': 123, 'neo4j_labels': ['Categories']}
        

LABELS

nameargumentsreturns
get_labelsself[str]
        Extract and return a list of ALL the Neo4j labels present in the database.
        No particular order should be expected.
        Note: to get the labels of a particular node, use get_node_labels()
        TODO: test when there are nodes that have multiple labels

        :return:    A list of strings
        
nameargumentsreturns
get_label_propertiesself, label:strlist
        Extract and return all the property (key) names used in nodes with the given label,
        sorted alphabetically

        :param label:   A string with the name of a node label
        :return:        A list of property names, sorted alphabetically
        

INDEXES

nameargumentsreturns
get_indexesselfpd.DataFrame
        Return all the database indexes, and some of their attributes,
        as a Pandas dataframe.

        EXAMPLE:
               labelsOrTypes              name          properties    type  uniqueness
             0    ["my_label"] "index_23b59623"    ["my_property"]   BTREE   NONUNIQUE
             1    ["L"]          "L.client_id"       ["client_id"]   BTREE      UNIQUE

        :return:        A (possibly-empty) Pandas dataframe
        
nameargumentsreturns
create_indexself, label :str, key :strbool
        Create a new database index, unless it already exists,
        to be applied to the specified label and key (property).
        The standard name given to the new index is of the form label.key
        EXAMPLE - to index nodes labeled "car" by their key "color", use:
                        create_index("car", "color")
                  This new index - if not already in existence - will be named "car.color"
        If an existing index entry contains a list of labels (or types) such as ["l1", "l2"] ,
        and a list of properties such as ["p1", "p2"] ,
        then the given pair (label, key) is checked against ("l1_l2", "p1_p2"), to decide whether it already exists.

        :param label:   A string with the node label to which the index is to be applied
        :param key:     A string with the key (property) name to which the index is to be applied
        :return:        True if a new index was created, or False otherwise
        
nameargumentsreturns
drop_indexself, name: strbool
        Get rid of the index with the given name

        :param name:    Name of the index to jettison
        :return:        True if successful or False otherwise (for example, if the index doesn't exist)
        
nameargumentsreturns
drop_all_indexesself, including_constraints=TrueNone
        Eliminate all the indexes in the database and, optionally, also get rid of all constraints

        :param including_constraints:   Flag indicating whether to also ditch all the constraints
        :return:                        None
        

CONSTRAINTS

nameargumentsreturns
get_constraintsselfpd.DataFrame
        Return all the database constraints, and some of their attributes,
        as a Pandas dataframe with 3 columns:
            name        EXAMPLE: "my_constraint"
            description EXAMPLE: "CONSTRAINT ON ( patient:patient ) ASSERT (patient.patient_id) IS UNIQUE"
            details     EXAMPLE: "Constraint( id=3, name='my_constraint', type='UNIQUENESS',
                                  schema=(:patient {patient_id}), ownedIndex=12 )"
        :return:  A (possibly-empty) Pandas dataframe
        
nameargumentsreturns
create_constraintself, label: str, key: str, type="UNIQUE", name=Nonebool
        Create a uniqueness constraint for a node property in the graph,
        unless a constraint with the standard name of the form `{label}.{key}.{type}` is already present
        Note: it also creates an index, and cannot be applied if an index already exists.
        EXAMPLE: create_constraint("patient", "patient_id")
        :param label:   A string with the node label to which the constraint is to be applied
        :param key:     A string with the key (property) name to which the constraint is to be applied
        :param type:    For now, the default "UNIQUE" is the only allowed option
        :param name:    Optional name to give to the new constraint; if not provided, a
                            standard name of the form `{label}.{key}.{type}` is used.  EXAMPLE: "patient.patient_id.UNIQUE"
        :return:        True if a new constraint was created, or False otherwise
        
nameargumentsreturns
drop_constraintself, name: strbool
        Eliminate the constraint with the specified name

        :param name:    Name of the constraint to eliminate
        :return:        True if successful or False otherwise (for example, if the constraint doesn't exist)
        
nameargumentsreturns
drop_all_constraintsselfNone
        Eliminate all the constraints in the database

        :return:    None
        

READ IN PANDAS

nameargumentsreturns
load_pandasself, df :Union[pd.DataFrame, pd.Series], labels :Union[str, List[str], Tuple[str]], merge_primary_key=None, merge_overwrite=False, rename=None, ignore_nan=True, max_chunk_size=10000[int]
        Load a Pandas Data Frame (or Series) into Neo4j.
        Each row is loaded as a separate node.

        Columns whose dtype is integer will appear as integer data types in the Neo4j nodes
            (however, if a NaN is present in a Pandas column, it'll automatically be transformed to float)

        Database indexes are added as needed, if the "merge_primary_key" argument is used.

        TODO: maybe save the Panda data frame's row number as an attribute of the Neo4j nodes, to ALWAYS have a primary key
              maybe allow a bulk auto-increment field
              maybe allow to link all nodes to a given existing one

        :param df:              A Pandas Data Frame (or Series) to import into Neo4j.
                                    If it's a Series, it's treated as a single column (named "value", if it lacks a name)

        :param labels:          A string, or list/tuple of strings - representing one or more Neo4j labels
                                    to use on all the newly-created nodes

        :param merge_primary_key: (OPTIONAL) A string with the name of the field that serves as a primary key
                                    If provided, new records will be merged (rather than added) if they already exist,
                                    as determined by that primary key
                                    TODO: to allow for list of primary_keys
        :param merge_overwrite: (OPTIONAL) Only applicable if "merge_primary_key" is set.
                                    If True then on merge the existing nodes will be completely overwritten with the new data,
                                    otherwise they will be updated with new information (keys that are not present
                                    in the df argument will be left unaltered)

        :param rename:          Optional dictionary to rename the Pandas dataframe's columns to
                                    EXAMPLE {"current_name": "name_we_want"}
        :param ignore_nan       If True, node properties created from columns of dtype float
                                    will only be set if they are not NaN.
                                    (Note: the moment a NaN is present, columns of integers in a dataframe
                                           will automatically become floats)
        :param max_chunk_size:  To limit the number of rows loaded at one time

        :return:                A (possibly-empty) list of the internal database ID's of the created nodes
        
nameargumentsreturns
pd_datetime_to_neo4j_datetimeself, df: pd.DataFramepd.DataFrame
        If any column in the given Pandas Data Frame is of dtype datetime or timedelta,
        clone the Data Frame and replace its entries with a Neo4j-friendly datetime type.

        If any change is needed, return a modified COPY of the dataframe;
        otherwise, return the original dataframe (no cloning done)

        Note: one must be vigilant against STRING columns with values that look like datetimes,
              but are not of that type!

        EXAMPLE: an entry such as pd.Timestamp('2023-01-01 00:00:00')
                 will become an object neo4j.time.DateTime(2023, 1, 1, 0, 0, 0, 0)

        :param df:  A Pandas data frame
        :return:    Either the same data frame, or a modified version of a clone of it
        

CSV IMPORT

nameargumentsreturns
import_csv_nodes_IN_PROGRESSself, filename :str, labels :Union[str, List[str], Tuple[str]], os="linux", headers=True, rename_fields=None, field_types=None, import_line_number=False, start_line_number=1, line_number_name="line number", link_to_node=Nonedict
        TODO: In-progress.  DO NOT USE.   Use load_pandas() instead
        
        Import data from a CSV file located on the same file system as the Neo4j database,
        and the records represent nodes to be created in the database.

        IMPORTANT:  the file to import MUST be located in the special folder NEO4J_HOME/import,
                    unless the Neo4j configuration file is first modified - and the database restarted!
                    A typical default location, when Neo4j is installed on Linux, is:  /var/lib/neo4j/import

        :param filename:    EXAMPLE in Linux:   "test.csv"
                            EXAMPLE in Windows: "C:/test.csv"
        :param labels:      A string, or list/tuple of strings, representing one or multiple Neo4j labels;
                                it's acceptable to be None
        :param os:          Either "linux" or "win"
        :param headers:         TODO: implement
        :param rename_fields:   TODO: implement
                                    Example: {"name": "Product Name", "n_parts": "Number of Parts"}
        :param field_types: (OPTIONAL) if not specified, all values get imported as STRINGS
                                TODO: implement
                                    Example: {"product_id": "int", "cost": "float"}
        :param import_line_number:
        :param start_line_number:
        :param line_number_name:
        :param link_to_node:    (OPTIONAL) If provided, all the newly-created nodes
                                    get linked to the specified existing node.
                                    Example: {"internal_id": 123, "name": "employed_by", "dir": "out"}

        :return:        A dictionary of statistics about the import
        

JSON IMPORT EXPORT

nameargumentsreturns
export_dbase_jsonself{}
        Export the entire Neo4j database as a JSON string.
        TODO: offer an option to automatically include today's date in the name of exported file

        IMPORTANT: APOC must be activated in the database, to use this function.
                   Otherwise it'll raise an Exception

        EXAMPLE:
        { 'nodes': 2,
          'relationships': 1,
          'properties': 6,
          'data': '[{"type":"node","id":"3","labels":["User"],"properties":{"name":"Adam","age":32,"male":true}},\n
                    {"type":"node","id":"4","labels":["User"],"properties":{"name":"Eve","age":18}},\n
                    {"id":"1","type":"relationship","label":"KNOWS","properties":{"since":2003},"start":{"id":"3","labels":["User"]},"end":{"id":"4","labels":["User"]}}\n
                   ]'
        }

        SIDE NOTE: the Neo4j Browser uses a slightly different format for NODES:
                {
                  "identity": 4,
                  "labels": [
                    "User"
                  ],
                  "properties": {
                    "name": "Eve",
                    "age": 18
                  }
                }
              and a substantially more different format for RELATIONSHIPS:
                {
                  "identity": 1,
                  "start": 3,
                  "end": 4,
                  "type": "KNOWS",
                  "properties": {
                    "since": 2003
                  }
                }

        :return:    A dictionary specifying the number of nodes exported ("nodes"),
                    the number of relationships ("relationships"),
                    and the number of properties ("properties"),
                    as well as a "data" field with the actual export as a JSON string
        
nameargumentsreturns
export_nodes_rels_jsonself, nodes_query="", rels_query=""{}
        Export the specified nodes, plus the specified relationships, as a JSON string.
        The default empty strings are taken to mean (respectively) ALL nodes/relationships.

        For details on the formats, see export_dbase_json()

        IMPORTANT:  APOC must be activated in the database for this function.
                    Otherwise it'll raise an Exception

        :param nodes_query: A Cypher query to identify the desired nodes (exclusive of RETURN statements)
                                    The dummy variable for the nodes must be "n"
                                    Use "" to request all nodes
                                    EXAMPLE: "MATCH (n) WHERE (n:CLASS OR n:PROPERTY)"
        :param rels_query:   A Cypher query to identify the desired relationships (exclusive of RETURN statements)
                                    The dummy variable for the relationships must be "r"
                                    Use "" to request all relationships (whether or not their end nodes are also exported)
                                    EXAMPLE: "MATCH ()-[r:HAS_PROPERTY]->()"

        :return:    A dictionary specifying the number of nodes exported,
                    the number of relationships, and the number of properties,
                    as well as a "data" field with the actual export as a JSON string

        
nameargumentsreturns
is_literalself, valuebool
        Return True if the given value represents a literal (in terms of database storage)

        :param value:
        :return:
        
nameargumentsreturns
import_jsonself, json_str: str, root_labels="import_root_label", parse_only=False, provenance=NoneList[int]
        Import the data specified by a JSON string into the database.

        CAUTION: A "postorder" approach is followed: create subtrees first (with recursive calls), then create root last;
        as a consequence, in case of failure mid-import, there's no top root, and there could be several fragments.
        A partial import might need to be manually deleted.
        TODO: maintain a list of all created nodes - so as to be able to delete them all in case of failure.

        :param json_str:    A JSON string representing the data to import
        :param root_labels: String, or list of strings, to be used as Neo4j labels for the root node(s)
        :param parse_only:  If True, the parsed data will NOT be added to the database
        :param provenance:  Optional string to store in a "source" attribute in the root node
                                (only used if the top-level JSON structure is an object, i.e. if there's a single root node)

        :return:            List of integer ID's (possibly empty), of the root node(s) created
        
nameargumentsreturns
create_nodes_from_python_dataself, python_data, root_labels: Union[str, List[str]], level=1List[int]
        Recursive function to add data from a JSON structure to the database, to create a tree:
        either a single node, or a root node with children.
        A "postorder" approach is followed: create subtrees first (with recursive calls), then create root last.

        If the data is a literal, first turn it into a dictionary using a key named "value".

        Return the Neo4j ID's of the root node(s)

        :param python_data: Python data to import.
                                The data can be a literal, or list, or dictionary
                                - and lists/dictionaries may be nested
        :param root_labels: String, or list of strings, to be used as Neo4j labels for the root node(s)
        :param level:       Recursion level (also used for debugging, to make the indentation more readable)
        :return:            List of integer Neo4j internal ID's (possibly empty), of the root node(s) created
        
nameargumentsreturns
dict_importerself, d: dict, labels, level: intint
        Import data from a Python dictionary.
        If the data is nested, it uses a recursive call to create_nodes_from_python_data()
        TODO: pytest

        :param d:       A Python dictionary with data to import
        :param labels:  String, or list of strings, to be used as Neo4j labels for the node
        :param level:   Integer with recursion level (used just to format debugging output)
        :return:        Integer with the internal database id of the newly-created (top-level) node
        
nameargumentsreturns
list_importerself, l: list, labels, level[int]
        Import data from a list.
        If the data is nested, it uses a recursive call to create_nodes_from_python_data()
        TODO: pytest

        :param l:       A list with data to import
        :param labels:  String, or list of strings, to be used as Neo4j labels for the node
        :param level:   Integer with recursion level (just used to format debugging output)
        :return:        List (possibly empty) of internal database id's of the newly-created nodes
        
nameargumentsreturns
import_json_dumpself, json_str: str, extended_validation = Truestr
        Used to import data from a database dump that was done with export_dbase_json() or export_nodes_rels_json().

        Import nodes and relationships into the database, as specified in the JSON code
        that was created by the earlier data dump.

        IMPORTANT: the internal id's of the nodes need to be shifted,
              because one cannot force the Neo4j internal id's to be any particular value...
              and, besides (if one is importing into an existing database), particular id's may already be taken.

        :param json_str:            A JSON string with the format specified under export_dbase_json()
        :param extended_validation: If True, an attempt is made to try to avoid partial imports,
                                        by running extended validations prior to importing
                                        (it will make a first pass thru the data, and hence take longer)

        :return:                    A status message with import details if successful;
                                        or raise an Exception if not.
                                        If an error does occur during import then the import is aborted -
                                        and the number of imported nodes & relationships is returned in the Exception raised.
        

DEBUGGING SUPPORT

nameargumentsreturns
debug_printself, info: str, trim=FalseNone
        If the class' property "debug" is set to True,
        print out the passed info string,
        optionally trimming it, if too long

        :param info:
        :param trim:
        :return:        None
        
nameargumentsreturns
debug_trimself, data, max_len = 150str
        Abridge the given data (first turning it into a string if needed), if excessively long,
        using ellipses " ..." for the omitted data.
        Return the abridged data.

        :param data:    String with data to possibly abridge
        :param max_len: Max number of characters to show from the data argument
        :return:        The (possibly) abridged text
        
nameargumentsreturns
debug_trim_printself, data, max_len = 150None
        Abridge the given data (first turning it into a string if needed),
        if it is excessively long; then print it

        :param data:    String with data to possibly abridge, and then print
        :param max_len: Max number of characters to show from the data argument
        :return:        None
        
nameargumentsreturns
indent_chooserself, level: intstr
        Create an indent based on a "level": handy for debugging recursive functions

        :param level:
        :return:
        
nameargumentsreturns
_debug_localselfstr
        Use to test the switch from a local to remote repository, for debugging

        :return:
        



Class NodeSpecs

    Helper class for the class "NeoAccess".
    Meant as a PRIVATE class for NeoAccess; not indicated for the end user.
    
    Validates and stores all the passed specifications (the "RAW match structure"),
    that are used to identify a node or group of nodes.

    Note:   NO database operation is actually performed

    IMPORTANT:  By our convention -
                    if internal_id is provided, all other conditions are DISREGARDED;
                    if it's missing, an implicit AND operation applies to all the specified conditions
                (Regardless, all the passed data is stored in this object)
    
nameargumentsreturns
__init__self, internal_id=None, labels=None, key_name=None, key_value=None, properties=None, clause=None, clause_dummy_name="n"
        ALL THE ARGUMENTS ARE OPTIONAL (no arguments at all means "match everything in the database")

        :param internal_id: An integer with the node's internal database ID.
                                If specified, it will lead to all the remaining arguments being DISREGARDED (though saved)

        :param labels:      A string (or list/tuple of strings) specifying one or more Neo4j labels.
                                (Note: blank spaces ARE allowed in the strings)
                                EXAMPLES:  "cars"
                                            ("cars", "powered vehicles")
                            Note that if multiple labels are given, then only nodes possessing ALL of them will be matched;
                            at present, there's no way to request an "OR" operation on labels

        :param key_name:    A string with the name of a node attribute; if provided, key_value must be present, too
        :param key_value:   The required value for the above key; if provided, key_name must be present, too
                                Note: no requirement for the key to be primary

        :param properties:  A (possibly-empty) dictionary of property key/values pairs, indicating a condition to match.
                                EXAMPLE: {"gender": "F", "age": 22}

        :param clause:      Either None, OR a (possibly empty) string containing a Cypher subquery,
                            OR a pair/list (string, dict) containing a Cypher subquery and the data-binding dictionary for it.
                            The Cypher subquery should refer to the node using the assigned dummy_node_name (by default, "n")
                                IMPORTANT:  in the dictionary, don't use keys of the form "n_par_i",
                                            where n is the dummy node name and i is an integer,
                                            or an Exception will be raised - those names are for internal use only
                                EXAMPLES:   "n.age < 25 AND n.income > 100000"
                                            ("n.weight < $max_weight", {"max_weight": 100})

        :param clause_dummy_name: A string with a name by which to refer to the nodes (by default, "n") in the clause;
                                only used if a `clause` argument is passed (in the absence of a clause, it's stored as None)
        
nameargumentsreturns
__str__selfstr
        Return a description of this object

        :return:
        



Class CypherMatch

    Helper class for the class "NeoAccess".
    Meant as a PRIVATE class for NeoAccess; not indicated for the end user.
    
    Objects of this class (sometimes referred to as a "PROCESSED match structures")
    are used to facilitate a user to specify a node in a wide variety of ways - and
    save those specifications, in a "pre-digested" way, to use as needed in Cypher queries.

    NO extra database operations are involved.
    
    They store the following 4 properties:

        1) "node":  a string, defining a node in a Cypher query, *excluding* the "MATCH" keyword
        2) "where": a string, defining the "WHERE" part of the subquery (*excluding* the "WHERE"), if applicable;
                    otherwise, a blank
        3) "data_binding":      a (possibly empty) data-binding dictionary
        4) "dummy_node_name":   a string used for the node name inside the Cypher query (by default, "n");
                                potentially relevant to the "node" and "where" values

        EXAMPLES:
            *   node: "(n)"
                    where: ""
                    data_binding: {}
                    dummy_node_name: "n"
            *   node: "(p :`person` )"
                    where: ""
                    data_binding: {}
                    dummy_node_name: "p"
            *   node: "(n  )"
                    where: "id(n) = 123"
                    data_binding: {}
                    dummy_node_name: "n"
            *   node: "(n :`car`:`surplus inventory` )"
                    where: ""
                    data_binding: {}
                    dummy_node_name: "n"
            *    node: "(n :`person` {`gender`: $n_par_1, `age`: $n_par_2})"
                    where: ""
                    data_binding: {"n_par_1": "F", "n_par_2": 22}
                    dummy_node_name: "n"
            *   node: "(n :`person` {`gender`: $n_par_1, `age`: $n_par_2})"
                    where: "n.income > 90000 OR n.state = 'CA'"
                    data_binding: {"n_par_1": "F", "n_par_2": 22}
                    dummy_node_name: "n"
            *   node: "(n :`person` {`gender`: $n_par_1, `age`: $n_par_2})"
                    where: "n.income > $min_income"
                    data_binding: {"n_par_1": "F", "n_par_2": 22, "min_income": 90000}
                    dummy_node_name: "n"
    
nameargumentsreturns
__init__self, node_specs, dummy_node_name_if_missing=None


        :param node_specs:                  Object of type "NodeSpecs"
        :param dummy_node_name_if_missing:  String that will be used ONLY if the object passed to in node_specs
                                                lacks that attribute
        
nameargumentsreturns
__str__selfstr
        Return a description of this object
        :return:
        
nameargumentsreturns
extract_nodeselfstr
        Return the node information to be used in composing Cypher queries

        :return:        A string with the node information, as needed by Cypher queries.  EXAMPLES:
                            "(n  )"
                            "(p :`person` )"
                            "(n :`car`:`surplus inventory` )"
                            "(n :`person` {`gender`: $n_par_1, `age`: $n_par_2})"
        
nameargumentsreturns
extract_dummy_nameselfstr
        Return the dummy node _name to be used in composing Cypher queries

        :return:        A string with the dummy node name (often "n" , or "to" , or "from")
        
nameargumentsreturns
unpack_matchselflist
        Return a list containing:
        [node, where, data_binding, dummy_node_name] ,
        for use in composing Cypher queries

        :return:                A list containing [node, where, data_binding, dummy_node_name]
        
nameargumentsreturns
extract_where_clauseselfstr
        Cleanup the WHERE clause, and prefix the "WHERE" keyword as needed

        TODO: new method to test
        :return:
        
nameargumentsreturns
assert_valid_structureselfNone
        Verify that the object is a valid one (i.e., correctly initialized); if not, raise an Exception
        TODO: NOT IN CURRENT USE.  Perhaps to phase out, or keep it but tighten its tests

        :return:        None
        



Class CypherUtils

    Helper STATIC class for the class "NeoAccess".
    Meant as a PRIVATE class for NeoAccess; not indicated for the end user.
    
nameargumentsreturns
process_match_structurecls, handle: Union[int, NodeSpecs], dummy_node_name="n", caller_method=NoneCypherMatch
        Accept either a valid internal database node ID, or a "NodeSpecs" object (a "raw match"),
        and turn it into a "CypherMatch" object (a "processed match")

        Note: no database operation is performed

        :param handle:          Either an integer with a valid internal database ID,
                                    or an object of type NodeSpecs
        :param dummy_node_name: A string that will be used inside a Cypher query, to refer to nodes
        :param caller_method:   (OPTIONAL) String with name of caller method, only used for more clear error messages

        :return:                A "CypherMatch" object (a "processed match"), used to identify a node,
                                    or group of nodes
        
nameargumentsreturns
check_match_compatibilitycls, match1: CypherMatch, match2: CypherMatchNone
        If the two given "CypherMatch" objects (i.e. PROCESSED match structures)
        are incompatible - in terms of collision in their dummy node names -
        raise an Exception.

        :param match1:  A CypherMatch" object to be used to identify a node, or group of nodes
        :param match2:  A CypherMatch" object to be used to identify a node, or group of nodes
        :return:        None
        
nameargumentsreturns
combined_wherecls, match1: CypherMatch, match2: CypherMatch, check_compatibility=Truestr
        Given the two "CypherMatch" objects (i.e. PROCESSED match structures),
        return the combined version of all their WHERE statements.
        Also prefix the WHERE keyword to the result (if appropriate);
        if there are no clauses, an empty string is returned (without the WHERE keyword.)
        For details, see prepare_where()

        :param match1:  A CypherMatch" object to be used to identify a node, or group of nodes
        :param match2:  A CypherMatch" object to be used to identify a node, or group of nodes
        :param check_compatibility: Use True if the individual matches are meant to refer to different nodes,
                                        and need to make sure there's no conflict in the dummy node names
        :return:        A string with the combined WHERE statement,
                            suitable for inclusion into a Cypher query (empty if there were no subclauses)
        
nameargumentsreturns
combined_data_bindingcls, match1: CypherMatch, match2: CypherMatchdict
        Given the two "CypherMatch" objects (i.e. PROCESSED match structures),
        return the combined version of all their data binding dictionaries.
        NOTE:  if the individual matches are meant to refer to different nodes,
                    need to first make sure there's no conflict in the dummy node names -
                    use check_match_compatibility() as needed.
                    In practice, combined_where() is typically run whenever combined_data_binding() is -
                    and the former can take care of checking for compatibility

        :param match1:  A CypherMatch" object to be used to identify a node, or group of nodes
        :param match2:  A CypherMatch" object to be used to identify a node, or group of nodes
        :return:        A (possibly empty) dict with the combined data binding dictionaries,
                            suitable for inclusion into a Cypher query
        
nameargumentsreturns
assert_valid_internal_idcls, internal_id: intNone
        Raise an Exception if the argument is not a valid Neo4j internal database ID

        :param internal_id: Alleged Neo4j internal database ID
        :return:            None
        
nameargumentsreturns
valid_internal_idcls, internal_id: intbool
        Return True if internal_id is a valid ID as used internally by the database
        (aka "Neo4j ID")

        :param internal_id: An alleged internal database ID
        :return:            True if internal_id is a valid internal database ID, or False otherwise
        
nameargumentsreturns
prepare_labelscls, labels :Union[str, List[str], Tuple[str]]str
        Turn the given string, or list/tuple of strings - representing one or more Neo4j labels - into a string
        suitable for inclusion in a Cypher query.
        Blanks ARE allowed in the names.
        EXAMPLES:
            "" or None          both give rise to    ""
            "client"            gives rise to   ":`client`"
            "my label"          gives rise to   ":`my label`"
            ["car", "vehicle"]  gives rise to   ":`car`:`vehicle`"

        :param labels:  A string, or list/tuple of strings, representing one or multiple Neo4j labels;
                            it's acceptable to be None
        :return:        A string suitable for inclusion in the node part of a Cypher query
        
nameargumentsreturns
prepare_wherecls, where_list: Union[str, list]str
        Given a WHERE clause, or list/tuple of them, combined them all into one -
        and also prefix the WHERE keyword to the result (if appropriate).
        The *combined* clauses of the WHERE statement are parentheses-enclosed, to protect against code injection

        EXAMPLES:   "" or "      " or [] or ("  ", "") all result in  ""
                    "n.name = 'Julian'" returns "WHERE (n.name = 'Julian')"
                        Likewise for ["n.name = 'Julian'"]
                    ("p.key1 = 123", "   ",  "p.key2 = 456") returns "WHERE (p.key1 = 123 AND p.key2 = 456)"

        :param where_list:  A string with a subclause, or list or tuple of subclauses,
                            suitable for insertion in a WHERE statement

        :return:            A string with the combined WHERE statement,
                            suitable for inclusion into a Cypher query (empty if there were no subclauses)
        
nameargumentsreturns
dict_to_cyphercls, data_dict: {}, prefix="par_"(str, {})
        Turn a Python dictionary (meant for specifying node or relationship attributes)
        into a string suitable for Cypher queries,
        plus its corresponding data-binding dictionary.

        EXAMPLE :
                {'cost': 65.99, 'item description': 'the "red" button'}

                will lead to the pair:
                    (
                        '{`cost`: $par_1, `item description`: $par_2}',
                        {'par_1': 65.99, 'par_2': 'the "red" button'}
                    )

        Note that backticks are used in the Cypher string to allow blanks in the key names.
        Consecutively-named dummy variables ($par_1, $par_2, etc) are used,
        instead of names based on the keys of the data dictionary (such as $cost),
        because the keys might contain blanks.

        SAMPLE USAGE:
            (cypher_properties, data_binding) = dict_to_cypher(data_dict)

        :param data_dict:   A Python dictionary
        :param prefix:      Optional prefix string for the data-binding dummy names (parameter tokens); handy to prevent conflict;
                                by default, "par_"

        :return:            A pair consisting of a string suitable for Cypher queries,
                                and a corresponding data-binding dictionary.
                            If the passed dictionary is empty or None,
                                the pair returned is ("", {})