Initial add IBM DB2 vector store .

[yiweiHeOSS]

This PR adds a new folder under libs for the DB2 vector store langchain connector code, i.e., libs/langchain-db2.
Other companies are doing the same things for their vector store (e.g. Google, AWS and Azure). Thanks for @ccurme to give me some guidance about where to put the code and @MateuszOssGit gave me the access to this repo .

Later, I will create a package on Pypi of the vector store connector, i.e. pip install langchain-db2. Note that it should have no impact on the current package pip install langchain-ibm. I just need a separate directory libs/langchain-db2 to put open-source code.

The major changes in this PR are in the files below:

• libs/langchain-db2/langchain_db2/db2vs.py - this is the source file of the db2 vector store
• libs/langchain-db2/docs/db2.ipynb - this is the Jupyter notebook for the demo and steps on how to use the db2 vector store
• libs/langchain-db2/tests/integration_tests/test_db2vs.py - This is the integration test

[shaikhq]

All changes look good to me.

[yiweiHeOSS]

Hi @MateuszOssGit @Mateusz-Switala @ccurme . Could you please review this PR? @shaikhq From our IBM DB2 team has approved it. It should be ready for you to review. If I need to do anything, please let me know. Thank you!

[ccurme]

It looks like CI was not enabled for the new library. I enabled it and added a placeholder test.

IMO, priorities are (in descending order):

• Get CI checks to pass (this will block merging)
• Get make integration_test to pass on CI (this will block release)
• Incorporate the standard test suite for vector stores (this is optional but nice to have)

[yiweiHeOSS]

@ccurme Thanks for your comments. I tried to fix most of the problems in the CI, but I still saw lint problems. Do we have a command to fix them automatically? Do you see other issues I need to fix? Thanks.

[ccurme]

@ccurme Thanks for your comments. I tried to fix most of the problems in the CI, but I still saw lint problems. Do we have a command to fix them automatically? Do you see other issues I need to fix? Thanks.

make format runs an auto-formatter but I don't think it will address the issues:

Error: langchain_db2/db2vs.py:296:89: E501 Line too long (99 > 88)
Error: langchain_db2/db2vs.py:334:89: E501 Line too long (158 > 88)
Error: langchain_db2/db2vs.py:455:9: F842 Local variable embedding_arr is annotated but never used

[MateuszOssGit]

I was able to successfully format all required changes to linting successfully passed.

@yiweiHeOSS @shaikhq Could you please check on my changes from PR above to verify if everything has been updated correctly with your requirements?

[yiweiHeOSS]

@ccurme @MateuszOssGit Thank you so much for your help! It looks like this PR is good to review again.

@MateuszOssGit, I only found one little thing that I changed. It looks like we can't print() in the testcase (not sure if we have a better way to print the result in test), but we still want to run the functions like _embed_documents() for testing purposes.

[yiweiHeOSS]

Hi @ccurme @MateuszOssGit, I resolved all the comments. If everything is good, can you please approve the PR so we can deliver it? Thanks. cc @Mateusz-Switala

[yiweiHeOSS]

@ccurme @MateuszOssGit :)

[yiweiHeOSS]

Thank you for reviewing it @Mateusz-Switala @MateuszOssGit

[yiweiHeOSS]

Hi @ccurme, could you approve this PR? Thanks.