Update reranker limits

[prasmussen15]

Important

Update reranker limits, add preliminary RRF reranker, and update version and dependencies.

• Behavior:
  • Updated reranker limits in edge_search, node_search, and community_search in search.py.
  • Added RRF as a preliminary reranker in node_search and edge_search.
  • Modified edge_bfs_search and node_bfs_search to include a limit parameter in search_utils.py.
• Renames:
  • Renamed input to input_data in create() in openai.py and voyage.py.
• Misc:
  • Updated version in pyproject.toml from 0.3.17 to 0.3.18.
  • Updated openai dependency to ^1.52.2 and ruff to ^0.7.1 in pyproject.toml.

^{This description was created by for 6bc0413. It will automatically update as commits are pushed.}

[ellipsis-dev]

👍 Looks good to me! Reviewed everything up to 898e139 in 18 seconds

More details

• Looked at 452 lines of code in 3 files
• Skipped 1 files when reviewing.
• Skipped posting 6 drafted comments based on config settings.

1. graphiti_core/search/search.py:57

• Draft comment:
  The indentation level for function parameters has been changed to 8 spaces. Consider reverting to 4 spaces for consistency with PEP 8 guidelines. This issue is present in multiple functions in this file.
• Reason this comment was not posted:
  Confidence changes required: 50%
  The indentation in the function definitions has been changed from 4 spaces to 8 spaces. This is inconsistent with the typical Python style guide (PEP 8) which recommends 4 spaces per indentation level. This change affects readability and consistency across the codebase.

2. graphiti_core/search/search_utils.py:64

• Draft comment:
  The indentation level for function parameters has been changed to 8 spaces. Consider reverting to 4 spaces for consistency with PEP 8 guidelines. This issue is present in multiple functions in this file.
• Reason this comment was not posted:
  Confidence changes required: 50%
  The indentation in the function definitions has been changed from 4 spaces to 8 spaces. This is inconsistent with the typical Python style guide (PEP 8) which recommends 4 spaces per indentation level. This change affects readability and consistency across the codebase.

3. graphiti_core/search/search_utils.py:698

• Draft comment:
  Using list(set(...)) to remove duplicates changes the order of elements, which might not be intended. Consider using an OrderedDict to maintain order while removing duplicates.
• Reason this comment was not posted:
  Confidence changes required: 50%
  In the maximal_marginal_relevance function, the use of list(set(...)) to remove duplicates changes the order of elements, which might not be intended. This could affect the ranking order of the results.

4. graphiti_core/search/search.py:57

• Draft comment:
  Function parameters should be aligned consistently. Adjust the indentation to align parameters with the opening parenthesis.
• Reason this comment was not posted:
  Confidence changes required: 80%
  The code has inconsistent indentation for function parameters, which is not idiomatic.

5. graphiti_core/search/search_utils.py:64

• Draft comment:
  Function parameters should be aligned consistently. Adjust the indentation to align parameters with the opening parenthesis.
• Reason this comment was not posted:
  Confidence changes required: 80%
  The code has inconsistent indentation for function parameters, which is not idiomatic.

6. pyproject.toml:1

• Draft comment:
  Ensure consistent indentation for all sections in the file.
• Reason this comment was not posted:
  Confidence changes required: 50%
  The code has inconsistent indentation for function parameters, which is not idiomatic.

Workflow ID: wflow_GpQ1nmY4a5Nw7Iu6

---

You can customize Ellipsis with 👍 / 👎 feedback, review rules, user-specific overrides, quiet mode, and more.

[ellipsis-dev]

👍 Looks good to me! Incremental review on 34b0b0b in 23 seconds

More details

• Looked at 362 lines of code in 2 files
• Skipped 0 files when reviewing.
• Skipped posting 4 drafted comments based on config settings.

1. graphiti_core/search/search.py:119

• Draft comment:
  Consider adding the RRF reranker to the community_search function for consistency with node_search and edge_search.
• Reason this comment was not posted:
  Confidence changes required: 50%
  The PR introduces a new RRF reranker in the node_search and edge_search functions. However, the community_search function does not include this new reranker, which might be an oversight if the intent was to apply RRF consistently across all search functions.

2. graphiti_core/search/search_utils.py:230

• Draft comment:
  Ensure the limit parameter is applied in the Cypher query to avoid large result sets.
• Reason this comment was not posted:
  Confidence changes required: 80%
  The PR modifies the edge_bfs_search and node_bfs_search functions to include a limit parameter. However, the limit parameter is not used in the Cypher query within edge_bfs_search, which could lead to performance issues if the result set is large.

3. graphiti_core/search/search_utils.py:621

• Draft comment:
  The rrf function is not immediately clear in its purpose. Consider adding a comment explaining its role in the code.
• Reason this comment was not posted:
  Confidence changes required: 50%
  The code is missing a comment explaining the purpose of the rrf function, which is not immediately clear from the code itself.

4. graphiti_core/search/search_utils.py:683

• Draft comment:
  The maximal_marginal_relevance function is not immediately clear in its purpose. Consider adding a comment explaining its role in the code.
• Reason this comment was not posted:
  Confidence changes required: 50%
  The code is missing a comment explaining the purpose of the maximal_marginal_relevance function, which is not immediately clear from the code itself.

Workflow ID: wflow_BxySWMF6aFFnJfnX

---

You can customize Ellipsis with 👍 / 👎 feedback, review rules, user-specific overrides, quiet mode, and more.

[ellipsis-dev]

👍 Looks good to me! Incremental review on 60c939e in 24 seconds

More details

• Looked at 31 lines of code in 2 files
• Skipped 0 files when reviewing.
• Skipped posting 3 drafted comments based on config settings.

1. graphiti_core/embedder/openai.py:45

• Draft comment:
  Renaming input to input_data avoids shadowing the built-in input function, which is a good practice. Ensure this change is consistently applied across the codebase.
• Reason this comment was not posted:
  Confidence changes required: 10%
  The change from input to input_data is consistent across both files and aligns with Python's best practices to avoid using built-in names as variable names.

2. graphiti_core/embedder/voyage.py:44

• Draft comment:
  Renaming input to input_data avoids shadowing the built-in input function, which is a good practice. Ensure this change is consistently applied across the codebase.
• Reason this comment was not posted:
  Confidence changes required: 10%
  The change from input to input_data is consistent across both files and aligns with Python's best practices to avoid using built-in names as variable names.

3. graphiti_core/embedder/openai.py:44

• Draft comment:
  Function create should have a more descriptive name to indicate its purpose, such as create_embedding. This applies to the same function in voyage.py.
• Reason this comment was not posted:
  Confidence changes required: 50%
  The code in both files is mostly fine, but there is a minor issue with the function naming pattern.

Workflow ID: wflow_WinEi0Ul6ILIytS1

---

You can customize Ellipsis with 👍 / 👎 feedback, review rules, user-specific overrides, quiet mode, and more.

[ellipsis-dev]

👍 Looks good to me! Incremental review on 6bc0413 in 10 seconds

More details

• Looked at 22 lines of code in 1 files
• Skipped 0 files when reviewing.
• Skipped posting 2 drafted comments based on config settings.

1. graphiti_core/embedder/voyage.py:53

• Draft comment:
  The check for empty strings in input_list is redundant since it's already done in the previous list comprehension. You can remove this line.
• Reason this comment was not posted:
  Confidence changes required: 50%
  The check for empty strings in input_list is redundant since it's already done in the previous list comprehension.

2. graphiti_core/embedder/voyage.py:49

• Draft comment:
  The list comprehension on line 49 and line 51 can be combined with the one on line 53 to avoid repetition and make the code more DRY.
• Reason this comment was not posted:
  Confidence changes required: 80%
  The code is mostly fine, but there are some improvements needed for idiomatic Python and DRY principles.

Workflow ID: wflow_gzXJmA7T77aMLeTe

---

You can customize Ellipsis with 👍 / 👎 feedback, review rules, user-specific overrides, quiet mode, and more.