feat: add kanon-2-reranker integration#1
Open
yasinkip wants to merge 4 commits into
Open
Conversation
There was a problem hiding this comment.
Pull request overview
Adds support for Isaacus “kanon-2-reranker” as a LangChain BaseDocumentCompressor, alongside docs and tests, and adjusts embeddings client construction to handle SecretStr API keys correctly.
Changes:
- Introduces
IsaacusRerank(sync + async) document reranking compressor. - Adds unit + integration tests for reranking behavior and initialization.
- Updates docs/README examples and fixes embeddings API key handling when passed as
SecretStr.
Reviewed changes
Copilot reviewed 8 out of 8 changed files in this pull request and generated 10 comments.
Show a summary per file
| File | Description |
|---|---|
langchain_isaacus/rerank.py |
New reranker integration implementing compress_documents / acompress_documents. |
langchain_isaacus/embeddings.py |
Adjusts client creation to unwrap SecretStr API keys. |
langchain_isaacus/__init__.py |
Exports IsaacusRerank from the package root. |
tests/unit_tests/test_rerank.py |
Adds unit tests for reranker initialization. |
tests/integration_tests/test_rerank.py |
Adds integration tests that call the live reranking API. |
docs/text_embedding.ipynb |
Updates embeddings notebook links/text formatting. |
docs/reranking.ipynb |
Adds reranking notebook with an end-to-end example. |
README.md |
Adds reranking usage snippet and updates embeddings snippet. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Comment on lines
+23
to
+25
| documents = [Document("foo bar"), Document("bar foo"), Document("foo")] | ||
| rerank= IsaacusRerank(model=MODEL) | ||
| output = rerank.compress_documents(documents, query) |
| @@ -0,0 +1,28 @@ | |||
| """Test embedding model integration.""" | |||
Comment on lines
+11
to
+13
| def test_initialization_kanon_2_reranker() -> None: | ||
| """Test embedding model initialization.""" | ||
| rr = IsaacusRerank(api_key=SecretStr("NOT_A_VALID_KEY"), model=MODEL) |
| "source": [ | ||
| "## Instantiation\n", | ||
| "\n", | ||
| "Now we can instantiate our model object and generate chat completions:" |
Comment on lines
13
to
22
| __all__ = [ | ||
| "ChatIsaacus", | ||
| "IsaacusVectorStore", | ||
| "IsaacusEmbeddings", | ||
| "IsaacusLoader", | ||
| "IsaacusRetriever", | ||
| "IsaacusRerank", | ||
| "IsaacusToolkit", | ||
| "IsaacusTool", | ||
| "__version__", |
| from langchain_core.documents import Document | ||
|
|
||
| rerankings = IsaacusRerank(model="kanon-2-reranker") | ||
| embeddings.compress_documents( |
| "# Show the ranked results\n", | ||
| "print(query)\n", | ||
| "for idx, doc in enumerate(reranked):\n", | ||
| " print(f\"Rank {idx+1} ({doc.metadata[\"score\"]:.4f}): {doc.page_content}\")" |
Comment on lines
+60
to
+66
| def rerank( | ||
| self, documents: Sequence[Union[Document, str]], query: str | ||
| ) -> list[dict[str, Any]]: | ||
| """Returns an ordered list of documents ordered by their relevance to | ||
| the query. | ||
| """ | ||
| docs = [self._document_to_str(doc) for doc in documents] |
Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
Collaborator
|
@copilot apply changes based on the comments in this thread |
There was a problem hiding this comment.
Pull request overview
Adds an Isaacus “kanon-2-reranker” integration to the langchain_isaacus package, alongside docs/examples and a small fix to the embeddings client initialization.
Changes:
- Introduces
IsaacusRerank(BaseDocumentCompressor) with sync/async reranking andtop_nsupport. - Fixes embeddings client construction to unwrap
SecretStrAPI keys before passing them to the Isaacus client. - Adds unit/integration tests and documentation (new reranking notebook + README updates).
Reviewed changes
Copilot reviewed 8 out of 8 changed files in this pull request and generated 9 comments.
Show a summary per file
| File | Description |
|---|---|
langchain_isaacus/rerank.py |
New reranker implementation (sync/async) that compresses documents by Isaacus relevance score. |
langchain_isaacus/embeddings.py |
Changes how API keys are passed to Isaacus clients (unwrap SecretStr). |
langchain_isaacus/__init__.py |
Exports IsaacusRerank at package level. |
tests/unit_tests/test_rerank.py |
Adds unit tests covering reranker initialization behavior. |
tests/integration_tests/test_rerank.py |
Adds integration tests for reranking behavior (sync/async, top_n). |
docs/text_embedding.ipynb |
Updates embedding notebook links/text and minor formatting. |
docs/reranking.ipynb |
Adds a new reranking “getting started” notebook. |
README.md |
Adds a reranking usage example and adjusts embeddings example. |
Comments suppressed due to low confidence (1)
langchain_isaacus/init.py:23
__all__lists several symbols (ChatIsaacus,IsaacusVectorStore,IsaacusLoader, etc.) that aren’t imported or defined anywhere in this package (the only modules present areembeddings.pyandrerank.py). As a result,from langchain_isaacus import ChatIsaacuswill raiseAttributeError. Either import/re-export these symbols here (possibly via lazy imports) or remove them from__all__to match what the package actually exposes.
__all__ = [
"ChatIsaacus",
"IsaacusVectorStore",
"IsaacusEmbeddings",
"IsaacusLoader",
"IsaacusRetriever",
"IsaacusRerank",
"IsaacusToolkit",
"IsaacusTool",
"__version__",
]
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Comment on lines
+34
to
+36
| rerankings = IsaacusRerank(model="kanon-2-reranker") | ||
| embeddings.compress_documents( | ||
| documents=[Document("The meaning of life is ...")], |
| "# Show the ranked results\n", | ||
| "print(query)\n", | ||
| "for idx, doc in enumerate(reranked):\n", | ||
| " print(f\"Rank {idx+1} ({doc.metadata[\"score\"]:.4f}): {doc.page_content}\")" |
Comment on lines
117
to
129
| def _get_api_key(self) -> str: | ||
| if isinstance(self.api_key, SecretStr): | ||
| api_key = self.api_key.get_secret_value() | ||
| else: | ||
| api_key = self.api_key | ||
|
|
||
| return api_key | ||
|
|
||
| def _get_client(self) -> Isaacus: | ||
| if not self._client: | ||
| self._client = Isaacus( | ||
| api_key=self.api_key, | ||
| api_key=self._get_api_key(), | ||
| base_url=self.base_url, |
Comment on lines
+13
to
+16
| query = "foo" | ||
| documents = [Document("foo bar")] | ||
| rerank = IsaacusRerank(model=MODEL) | ||
| output = rerank.compress_documents(documents, query) |
Comment on lines
+31
to
+34
| query = "foo" | ||
| documents = [Document("foo bar"), Document("bar foo"), Document("foo")] | ||
| rerank = IsaacusRerank(model=MODEL) | ||
| output = await rerank.acompress_documents(documents, query) |
Comment on lines
+1
to
+14
| """Test embedding model integration.""" | ||
|
|
||
| from langchain_core.documents import BaseDocumentCompressor | ||
| from pydantic import SecretStr | ||
|
|
||
| from langchain_isaacus import IsaacusRerank | ||
|
|
||
| MODEL = "kanon-2-reranker" | ||
|
|
||
|
|
||
| def test_initialization_kanon_2_reranker() -> None: | ||
| """Test embedding model initialization.""" | ||
| rr = IsaacusRerank(api_key=SecretStr("NOT_A_VALID_KEY"), model=MODEL) | ||
| assert isinstance(rr, BaseDocumentCompressor) |
| "source": [ | ||
| "## Instantiation\n", | ||
| "\n", | ||
| "Now we can instantiate our model object and generate chat completions:" |
Comment on lines
+22
to
+25
| query = "foo" | ||
| documents = [Document("foo bar"), Document("bar foo"), Document("foo")] | ||
| rerank = IsaacusRerank(model=MODEL) | ||
| output = rerank.compress_documents(documents, query) |
Comment on lines
+40
to
+43
| query = "foo" | ||
| documents = [Document("foo bar"), Document("bar foo"), Document("foo")] | ||
| rerank = IsaacusRerank(model=MODEL, top_n=2) | ||
| output = rerank.compress_documents(documents, query) |
Collaborator
|
@copilot apply changes based on the comments in this thread |
Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Kanon 2 Reranker Integration
Features:
Documentation:
Fixes: