langchain - ✅(Solved) Fix WeaviateVectorStore fails silently on index [4 pull requests, 2 comments, 2 participants]

langchain2026-03-05 10:16:57

ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

GitHub issue URL

Helpful · Quick feedback

GitHub stats

langchain-ai/langchain#35572•Fetched 2026-04-08 00:25:30

View on GitHub

Comments

Participants

Timeline

Reactions

Author

xQsM3

Participants

JiwaniZakir

xQsM3

Timeline (top)

referenced ×5cross-referenced ×4commented ×2labeled ×2

the add_documents / add_texts in WeaviateVectorStore logs errors during indexing but still returns a valid list of chunk IDs, giving the impression that indexing was successfull especially in fastapi apps where the langchain logger is not attached to the runtime. there should be imho a way at least something like failed_objects, success_objects = store.add_documents(), so that it becomes clear for the implementation side of add_documents which chunks failed:

Error Message

logger.error(err_message)

Error Message and Stack Trace (if applicable)

Root Cause

Fix Action

Fix / Workaround

This is a bug, not a usage question.
I added a clear and descriptive title that summarizes this issue.
I used the GitHub search to find a similar question and didn't find it.
I am sure that this is a bug in LangChain rather than my code.
The bug is not resolved by updating to the latest stable version of LangChain (or the specific integration package).
This is not related to the langchain-community package.
I posted a self-contained, minimal, reproducible example. A maintainer can copy it and run it AS IS.

Other Dependencies

aiohttp: 3.12.14 anthropic: 0.75.0 blockbuster: 1.5.25 click: 8.1.8 cloudpickle: 3.1.1 cryptography: 44.0.1 dataclasses-json: 0.6.7 grpcio: 1.76.0 grpcio-health-checking: 1.76.0 grpcio-tools: 1.75.1 httpx: 0.28.1 httpx-sse: 0.4.0 jsonpatch: 1.33 jsonschema-rs: 0.29.1 langgraph: 1.0.5 langgraph-checkpoint: 3.0.1 numpy: 2.2.6 openai: 2.14.0 opentelemetry-api: 1.39.0 opentelemetry-exporter-otlp-proto-http: 1.39.0 opentelemetry-sdk: 1.39.0 orjson: 3.11.5 packaging: 25.0 protobuf: 6.33.2 pydantic: 2.12.5 pydantic-settings: 2.12.0 pyjwt: 2.10.1 pytest: 8.3.4 python-dotenv: 1.1.0 pyyaml: 6.0.3 PyYAML: 6.0.3 requests: 2.32.5 requests-toolbelt: 1.0.0 rich: 13.9.4 simsimd: 6.5.12 sqlalchemy: 2.0.40 SQLAlchemy: 2.0.40 sse-starlette: 2.1.3 starlette: 0.46.2 structlog: 25.2.0 tenacity: 9.1.2 tiktoken: 0.8.0 truststore: 0.10.4 typing-extensions: 4.15.0 uuid-utils: 0.12.0 uvicorn: 0.34.0 watchfiles: 1.0.4 weaviate-client: 4.14.0 zstandard: 0.23.0 (venv) x@LAPTOP-97M072GO:~/Repos/chain-pipes$

PR fix notes

PR #260: fix(weaviate): raise on batch indexing failures in add_texts

Repository: langchain-ai/langchain-weaviate
Author: JasonCZMeng
State: open | merged: False
Link: https://github.com/langchain-ai/langchain-weaviate/pull/260

Description (problem / solution / changelog)

Summary

add_texts() now raises a ValueError when Weaviate batch indexing fails, instead of silently logging errors and returning all IDs as if everything succeeded.

Fixes langchain-ai/langchain#35572

Problem

WeaviateVectorStore.add_texts() calls self._client.batch.failed_objects after batch indexing, logs any failures, but still returns the full list of IDs — giving callers no indication that some objects weren't actually indexed. This is especially problematic in FastAPI apps where the LangChain logger may not be attached.

Changes

libs/weaviate/langchain_weaviate/vectorstores.py

After logging failures, filter failed UUIDs out of the returned ids list
Raise ValueError with a descriptive message: count of failures, total attempted, and first error
Attach successful_ids attribute to the exception for partial recovery (getattr(e, 'successful_ids', []))

libs/weaviate/tests/unit_tests/test_vectorstores_unit.py

Added test_add_texts_raises_value_error_on_batch_failures() — mocks one failed object out of two, asserts ValueError is raised with correct message and successful_ids contains only the non-failed ID

Backward Compatibility

This is a behavior change from silent failure to raising. However, silent data loss on batch indexing is a bug, not a feature — callers relying on the previous behavior were already losing data without knowing it.

Changed files

libs/weaviate/langchain_weaviate/vectorstores.py (modified, +11/-0)
libs/weaviate/tests/unit_tests/test_vectorstores_unit.py (modified, +40/-0)

PR #261: fix: raise ValueError when batch operations fail in add_texts

Repository: langchain-ai/langchain-weaviate
Author: xin700
State: open | merged: False
Link: https://github.com/langchain-ai/langchain-weaviate/pull/261

Description (problem / solution / changelog)

Description

Fixes langchain-ai/langchain#35572

Currently, WeaviateVectorStore.add_texts() only logs errors when batch operations fail, but still returns all IDs (including failed ones). This causes silent failures, especially in FastAPI apps where logs aren't visible.

Changes

Modified add_texts() to raise ValueError with detailed error messages when any objects fail to index
Added unit test to verify the new behavior

Impact

Breaking change: Code that previously ignored failures will now get exceptions Benefit: Developers will immediately know when indexing fails

Example

Before: Silent failure After: Raises ValueError with details

Changed files

libs/weaviate/langchain_weaviate/vectorstores.py (modified, +12/-5)
libs/weaviate/tests/unit_tests/test_vectorstores_unit.py (modified, +25/-0)

PR #35632: fix: WeaviateVectorStore fails silently on index

Repository: langchain-ai/langchain
Author: JiwaniZakir
State: closed | merged: False
Link: https://github.com/langchain-ai/langchain/pull/35632

Description (problem / solution / changelog)

Summary

This addresses #35572.

What changed

Implemented a fix based on the issue description. See the diff for specifics.

Testing

Verified against the existing test suite
Checked that the fix addresses the reported behavior

Changed files

libs/partners/weaviate/langchain_weaviate/__init__.py (added, +5/-0)
libs/partners/weaviate/langchain_weaviate/_math.py (added, +138/-0)
libs/partners/weaviate/langchain_weaviate/utils.py (added, +52/-0)
libs/partners/weaviate/langchain_weaviate/vectorstores.py (added, +563/-0)
libs/partners/weaviate/pyproject.toml (added, +107/-0)
libs/partners/weaviate/tests/__init__.py (added, +0/-0)
libs/partners/weaviate/tests/unit_tests/__init__.py (added, +0/-0)
libs/partners/weaviate/tests/unit_tests/test_vectorstores_unit.py (added, +307/-0)

PR #262: fix: raise on WeaviateVectorStore batch indexing failures

Repository: langchain-ai/langchain-weaviate
Author: s-zx
State: open | merged: False
Link: https://github.com/langchain-ai/langchain-weaviate/pull/262

Description (problem / solution / changelog)

Summary

Raise ValueError when Weaviate batch indexing fails instead of silently returning IDs. Fixes langchain-ai/langchain#35572.

Root Cause

add_texts logged errors via logger.error but still returned all IDs, giving false success in FastAPI apps where logs are not visible.

Fix

Check failed_objs and raise ValueError with details before returning.

Fixes langchain-ai/langchain#35572

Changed files

libs/weaviate/langchain_weaviate/vectorstores.py (modified, +4/-6)

Code Example

def add_texts(
        self,
        texts: Iterable[str],
        metadatas: Optional[List[dict]] = None,
        tenant: Optional[str] = None,
        **kwargs: Any,
    ) -> List[str]:
        """Upload texts with metadata (properties) to Weaviate."""
        from weaviate.util import get_valid_uuid  # type: ignore

        if tenant and not self._does_tenant_exist(tenant):
            logger.info(
                f"Tenant {tenant} does not exist in index {self._index_name}. "
                "Creating tenant."
            )
            tenant_objs = [weaviate.classes.tenants.Tenant(name=tenant)]
            self._collection.tenants.create(tenants=tenant_objs)

        ids = []
        embeddings: Optional[List[List[float]]] = None
        if self._embedding:
            embeddings = self._embedding.embed_documents(list(texts))

        with self._client.batch.dynamic() as batch:
            for i, text in enumerate(texts):
                data_properties = {self._text_key: text}
                if metadatas is not None:
                    for key, val in metadatas[i].items():
                        data_properties[key] = _json_serializable(val)

                # Allow for ids (consistent w/ other methods)
                # # Or uuids (backwards compatible w/ existing arg)
                # If the UUID of one of the objects already exists
                # then the existing object will be replaced by the new object.
                _id = get_valid_uuid(uuid4())
                if "uuids" in kwargs:
                    _id = kwargs["uuids"][i]
                elif "ids" in kwargs:
                    _id = kwargs["ids"][i]

                batch.add_object(
                    collection=self._index_name,
                    properties=data_properties,
                    uuid=_id,
                    vector=embeddings[i] if embeddings else None,
                    tenant=tenant,
                )

                ids.append(_id)

        failed_objs = self._client.batch.failed_objects
        for obj in failed_objs:
            err_message = (
                f"Failed to add object: {obj.original_uuid}\nReason: {obj.message}"
            )

            logger.error(err_message)

        return ids

---

RAW_BUFFERClick to expand / collapse

Checked other resources

This is a bug, not a usage question.
I added a clear and descriptive title that summarizes this issue.
I used the GitHub search to find a similar question and didn't find it.
I am sure that this is a bug in LangChain rather than my code.
The bug is not resolved by updating to the latest stable version of LangChain (or the specific integration package).
This is not related to the langchain-community package.
I posted a self-contained, minimal, reproducible example. A maintainer can copy it and run it AS IS.

Package (Required)

Related Issues / PRs

No response

Reproduction Steps / Example Code (Python)

def add_texts(
        self,
        texts: Iterable[str],
        metadatas: Optional[List[dict]] = None,
        tenant: Optional[str] = None,
        **kwargs: Any,
    ) -> List[str]:
        """Upload texts with metadata (properties) to Weaviate."""
        from weaviate.util import get_valid_uuid  # type: ignore

        if tenant and not self._does_tenant_exist(tenant):
            logger.info(
                f"Tenant {tenant} does not exist in index {self._index_name}. "
                "Creating tenant."
            )
            tenant_objs = [weaviate.classes.tenants.Tenant(name=tenant)]
            self._collection.tenants.create(tenants=tenant_objs)

        ids = []
        embeddings: Optional[List[List[float]]] = None
        if self._embedding:
            embeddings = self._embedding.embed_documents(list(texts))

        with self._client.batch.dynamic() as batch:
            for i, text in enumerate(texts):
                data_properties = {self._text_key: text}
                if metadatas is not None:
                    for key, val in metadatas[i].items():
                        data_properties[key] = _json_serializable(val)

                # Allow for ids (consistent w/ other methods)
                # # Or uuids (backwards compatible w/ existing arg)
                # If the UUID of one of the objects already exists
                # then the existing object will be replaced by the new object.
                _id = get_valid_uuid(uuid4())
                if "uuids" in kwargs:
                    _id = kwargs["uuids"][i]
                elif "ids" in kwargs:
                    _id = kwargs["ids"][i]

                batch.add_object(
                    collection=self._index_name,
                    properties=data_properties,
                    uuid=_id,
                    vector=embeddings[i] if embeddings else None,
                    tenant=tenant,
                )

                ids.append(_id)

        failed_objs = self._client.batch.failed_objects
        for obj in failed_objs:
            err_message = (
                f"Failed to add object: {obj.original_uuid}\nReason: {obj.message}"
            )

            logger.error(err_message)

        return ids

Error Message and Stack Trace (if applicable)

Description

System Info

python -m langchain_core.sys_info

System Information

OS: Linux OS Version: #1 SMP Tue Nov 5 00:21:55 UTC 2024 Python Version: 3.12.7 | packaged by Anaconda, Inc. | (main, Oct 4 2024, 13:27:36) [GCC 11.2.0]

Package Information

langchain_core: 1.2.6 langchain: 1.2.0 langchain_community: 0.4.1 langsmith: 0.3.45 langchain_anthropic: 1.0.0 langchain_classic: 1.0.1 langchain_openai: 1.1.6 langchain_text_splitters: 1.1.0 langchain_weaviate: 0.0.6 langgraph_api: 0.6.24 langgraph_cli: 0.4.11 langgraph_runtime_inmem: 0.21.1 langgraph_sdk: 0.3.1

Optional packages not installed

langserve

Other Dependencies

aiohttp: 3.12.14 anthropic: 0.75.0 blockbuster: 1.5.25 click: 8.1.8 cloudpickle: 3.1.1 cryptography: 44.0.1 dataclasses-json: 0.6.7 grpcio: 1.76.0 grpcio-health-checking: 1.76.0 grpcio-tools: 1.75.1 httpx: 0.28.1 httpx-sse: 0.4.0 jsonpatch: 1.33 jsonschema-rs: 0.29.1 langgraph: 1.0.5 langgraph-checkpoint: 3.0.1 numpy: 2.2.6 openai: 2.14.0 opentelemetry-api: 1.39.0 opentelemetry-exporter-otlp-proto-http: 1.39.0 opentelemetry-sdk: 1.39.0 orjson: 3.11.5 packaging: 25.0 protobuf: 6.33.2 pydantic: 2.12.5 pydantic-settings: 2.12.0 pyjwt: 2.10.1 pytest: 8.3.4 python-dotenv: 1.1.0 pyyaml: 6.0.3 PyYAML: 6.0.3 requests: 2.32.5 requests-toolbelt: 1.0.0 rich: 13.9.4 simsimd: 6.5.12 sqlalchemy: 2.0.40 SQLAlchemy: 2.0.40 sse-starlette: 2.1.3 starlette: 0.46.2 structlog: 25.2.0 tenacity: 9.1.2 tiktoken: 0.8.0 truststore: 0.10.4 typing-extensions: 4.15.0 uuid-utils: 0.12.0 uvicorn: 0.34.0 watchfiles: 1.0.4 weaviate-client: 4.14.0 zstandard: 0.23.0 (venv) x@LAPTOP-97M072GO:~/Repos/chain-pipes$

extent analysis

Problem Summary

The add_texts method in WeaviateVectorStore logs errors during indexing but still returns a valid list of chunk IDs, making it difficult to determine which chunks failed.

Root Cause Analysis

The root cause is likely due to the fact that the add_texts method does not properly handle failed objects. It logs the errors but continues to process the remaining objects, resulting in a valid list of chunk IDs being returned.

Fix Plan

To fix this issue, we need to modify the add_texts method to properly handle failed objects. Here are the steps to fix this issue:

Step 1: Modify the `add_texts` method to return failed objects

def add_texts(
        self,
        texts: Iterable[str],
        metadatas: Optional[List[dict]] = None,
        tenant: Optional[str] = None,
        **kwargs: Any,
    ) -> Tuple[List[str], List[FailedObject]]:
    ...
    failed_objs = self._client.batch.failed_objects
    return ids, failed_objs

Step 2: Update the caller to handle failed objects

failed_objects, success_objects = store.add_texts(texts)
for failed_obj in failed_objects:
    logger.error(f"Failed to add object: {failed_obj.original_uuid}\nReason: {failed_obj.message}")

Step 3: Update the logging to include failed objects

logger.info(f"Added {len(success_objects)} objects successfully")
logger.error(f"Failed to add {len(failed_objects)} objects: {failed_objects}")

Verification

To verify that the fix worked, you can run the add_texts method with a sample dataset and check the logs for any failed objects. You can also add additional logging statements to verify that the failed objects are being properly handled.

Extra Tips

Make sure to update the logging configuration to include the failed objects.
Consider

Vote matrix · Quick signals

Works

Did the solution work? Tap to confirm.

Easy Fix

Was it a quick fix?

Time Saver

Did it save you time?

Blocking

Was it severely blocking?

Common Issue

Are others likely hitting this too?

Flaky / Intermittent

Is it intermittent?

Verified / Reproducible

Can you reproduce it reliably?

#api #ssr #installation #tensor shape #autograd error #GPU setup #container setup #orchestration issue

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

langchain - ✅(Solved) Fix WeaviateVectorStore fails silently on index [4 pull requests, 2 comments, 2 participants]

Recommended Tools

GitHub issue graph ai analysis

Error Message

Error Message and Stack Trace (if applicable)

Root Cause

Fix Action

Fix / Workaround

Other Dependencies

PR fix notes

PR #260: fix(weaviate): raise on batch indexing failures in add_texts

Description (problem / solution / changelog)

Summary

Problem

Changes

Backward Compatibility

Changed files

PR #261: fix: raise ValueError when batch operations fail in add_texts

Description (problem / solution / changelog)

Description

Changes

Impact

Example

Changed files

PR #35632: fix: WeaviateVectorStore fails silently on index

Description (problem / solution / changelog)

Summary

What changed

Testing

Changed files

PR #262: fix: raise on WeaviateVectorStore batch indexing failures

Description (problem / solution / changelog)

Summary

Root Cause

Fix

Changed files

Code Example

Checked other resources

Package (Required)

Related Issues / PRs

Reproduction Steps / Example Code (Python)

Error Message and Stack Trace (if applicable)

Description

System Info

System Information

Package Information

Optional packages not installed

Other Dependencies

extent analysis

Problem Summary

Root Cause Analysis

Fix Plan

Step 1: Modify the add_texts method to return failed objects

Step 2: Update the caller to handle failed objects

Step 3: Update the logging to include failed objects

Verification

Extra Tips

Still need to ship something?

RELATED_DISCOVERY

TRENDING

Step 1: Modify the `add_texts` method to return failed objects