v0.14.0

Fixed

• Fix bug where patching previous valid nodes into an invalid data model would throw an error

Changed

• Add comma separator for numbers in GraphEDA report
• Remove deprecated IngestionGenerator class
• Update README to include GraphEDA module
• Remove images/
• Mild refactor of Property.type validation

Added

• Add relType to relationship properties section of GraphEDA module
• Add arguments to GraphEDA.run() to allow for more flexibility in chosen methods and parameters
• Add arguments to GraphEDA.node_degrees to allow for better result filtering
• Add arguments to GraphEDA.run() and GraphEDA.create_eda_report() to control directly returning result

What's Changed

• Update init.py by @a-s-g93 in #139
• Improve grapheda formatting by @a-s-g93 in #140
• improve run() flexibility by @a-s-g93 in #141
• Fix modeling bug by @a-s-g93 in #143
• Clean up by @a-s-g93 in #144
• Clean up by @a-s-g93 in #145
• prep v0.14.0 by @a-s-g93 in #146

Full Changelog: v0.13.1...v0.14.0

v0.13.1

Fixed

• Remove dotenv package dependency from Neo4jGraph and GraphEDA

Full Changelog: v0.13.0...v0.13.1

v0.13.0

Added

• GraphEDA module that gathers analytics and provides a report on the specified database
• gds_version property to Neo4jGraph
• Examples demonstrating how to use the new GraphEDA module

What's Changed

• add version by @a-s-g93 in #134
• Grapheda by @a-s-g93 in #135
• prep v0.13 release by @a-s-g93 in #136

Full Changelog: v0.12.0...v0.13.0

v0.12.0

Fixed

• Fix bug where use_cases arg in UserInput not accounted for.

Changed

• Change initial DataModel generation logic to first generate nodes, then generate relationships
• Updated examples
• remove use_yaml_data_model arg from DataModel generation methods, as it is no longer relevant
• Update DataModel, Node, Relationship and Property validation logic to better utilize Pydantic library
• Update Instructor dependency to v1.5.2
• Simplify LLM retry logic by better utilizing Instructor library
• Update printed results and progress of DataModel generating methods to be prettier
• Update Graphviz visual from DataModel.visualize()

Added

• Add DataModel validator to check for parallel relationships
• Add allow_parallel_relationships arg to DataModel generating methods
• Add apply_neo4j_naming_conventions arg to DataModel generating methods
• Add get_schema() to DataModel, Node, Relationship and Property to output a text version of schema

What's Changed

• fix bug in user input where use_case arg not accounted for by @a-s-g93 in #128
• Data modeling improvements by @a-s-g93 in #129
• create text schema output for data model, improve graphviz visual by @a-s-g93 in #130
• clean comments, update some docstrings by @a-s-g93 in #131
• Clean repo by @a-s-g93 in #132
• update docs by @a-s-g93 in #133

Full Changelog: v0.11.0...v0.12.0

v0.11.0

0.11.0

Changed

• Removed kwargs from LLM classes and replaced with llm_init_params to provide parameters that should be passed to the LLM constructor.

Added

• Azure OpenAI integrations for Discovery and Data Modeling: AzureOpenAIDiscoveryLLM and AzureOpenAIDataModelingLLM
• Added example that used Azure OpenAI models

What's Changed

• add Azure OpenAI integration by @a-s-g93 in #125
• Update neo4jRunway_e2e.ipynb by @a-s-g93 in #127

Full Changelog: v0.10.0...v0.11.0

v0.10.0

Fixed

• Initial data modeling takes into account entity pool generation step
• Fixed bug where some filepaths break saving content to file

Changed

• All data input to Discovery is converted to TableCollection class
• Discovery generated content is contained in Table on discovery_content attribute
• Discovery file output is handled on Table and TableCollection classes instead of Discovery class
• Data Modeling prompts have been significantly refactored and organized into a consistent format

Added

• Add verbose arg to PyIngest to suppress progress printing
• Multi file support for Discovery process (EXPERIMENTAL)
  • parameters in discovery.run() and discovery.run_async() to control process
  • async capability to improve response time for multi file
• Multi file support for Data Modeling process (EXPERIMENTAL)
• Cypher generation support for relationships spanning across different files

What's Changed

• update docs examples by @a-s-g93 in #114
• Add Example Notebooks by @djb91 in #118
• Updated README files by @djb91 in #119
• multi file input for discovery and data modeling - experimental by @a-s-g93 in #123
• prep v0.10 by @a-s-g93 in #124

Full Changelog: v0.9.1...v0.10.0

v0.9.1

Full Changelog: v0.9.0...v0.9.1

Fix bug in LoadCSVCodeGenerator

Unit tests for LoadCSVCodeGenerator and StandardCypherCodeGenerator classes

v0.9.0

Full Changelog: v0.8.2...v0.9.0

Fixed

pre-commit hooks now work properly and utilize ruff and mypy instead of black

Changed

Refactor LLM class into base class with DiscoveryLLM and DataModelingLLM child classes for each LLM integration

Restructure and refactor tests directory to mirror the neo4j_runway package structure

Implement pytest to handle testing

Added

Implement GitHub actions to automate unit and integration testing on PR

Implement GitHub actions to automate ruff and mypy checks on PR

Implement strict style guides utilizing ruff and mypy

Add Neo4jGraph module to handle database connections. This will be used in a future EDA module.

Add changelog to track changes better.

Use cases can now be added to the UserInput object to be considered during Discovery and Graph Data Modeling

Refactor discovery and data modeling prompts

Added additional validation checks to DataModel

Implement chain-of-thought reasoning for initial data model generation. This improves responses by:

Multi-hop traversals generated more reliably when appropriate

Constraints generated more reliably

v0.8.2

Patch Instructor version to 1.2.0 to fix errors from 1.3.7

Full Changelog: v0.8.1...v0.8.2

v0.8.1

What's Changed

• patch by @a-s-g93 in #95

• v0.8.1 by @a-s-g93 in #97

• Fix issue where sometimes error fix recommendations for modeling would provide conflicting solutions.

Full Changelog: v0.8.0...v0.8.1