Fix base CI issues and the defaults in .env.example (#24)

* Add the LLM_PROVIDER in example

* fix the assert in testing recall

* trial to fix failing client tests

NotImplementedError: Cannot copy out of meta tensor; no data! Please use torch.nn.Module.to_empty() instead of torch.nn.Module.to() when moving module from meta to a different device.

* lock the sentence transformer packages to align with the breaking changes around lazy tensor loading

* Add the LLM_PROVIDER in example

* fix the assert in testing recall

* trial to fix failing client tests

* pre-cache the model so CI doesn't need workarounds

* remove assert that is a race condition

The test was checking that the bank count increased, but with parallel tests (-n 8), other tests can delete their banks while this test is running, causing a race condition. The important assertion is assert test_bank_id in final_banks - which verifies the bank was actually created.

* add debug to figure out why docker build fails sometimes

* use the CPU only version of pytorch to avoid pulling cuda libraries

* add best match strategy to uv

* change the example openai model

.env.example

@@ -2,8 +2,9 @@
 # Copy this file to .env and fill in your values
 
 # LLM Configuration (Required)
+HINDSIGHT_API_LLM_PROVIDER=openai
 HINDSIGHT_API_LLM_API_KEY=your-api-key-here
-HINDSIGHT_API_LLM_MODEL=gpt-4o-mini
+HINDSIGHT_API_LLM_MODEL=o3-mini
 HINDSIGHT_API_LLM_BASE_URL=https://api.openai.com/v1
 
 # API Configuration (Optional)

.github/workflows/test.yml

@@ -108,6 +108,8 @@ jobs:
       HINDSIGHT_API_LLM_API_KEY: ${{ secrets.GROQ_API_KEY }}
       HINDSIGHT_API_LLM_MODEL: openai/gpt-oss-20b
       GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+      # Prefer CPU-only PyTorch in CI (but keep PyPI for everything else)
+      UV_INDEX: pytorch=https://download.pytorch.org/whl/cpu
 
     steps:
     - uses: actions/checkout@v4
@@ -132,7 +134,27 @@ jobs:
 
     - name: Install dependencies
       working-directory: ./hindsight-api
-      run: uv sync --extra test
+      run: uv sync --extra test --no-install-project --index-strategy unsafe-best-match
+
+    - name: Cache HuggingFace models
+      uses: actions/cache@v4
+      with:
+        path: ~/.cache/huggingface
+        key: ${{ runner.os }}-huggingface-${{ hashFiles('hindsight-api/pyproject.toml') }}
+        restore-keys: |
+          ${{ runner.os }}-huggingface-
+
+    - name: Pre-download models
+      working-directory: ./hindsight-api
+      run: |
+        uv run python -c "
+        from sentence_transformers import SentenceTransformer, CrossEncoder
+        print('Downloading embedding model...')
+        SentenceTransformer('BAAI/bge-small-en-v1.5')
+        print('Downloading cross-encoder model...')
+        CrossEncoder('cross-encoder/ms-marco-MiniLM-L-6-v2')
+        print('Models downloaded successfully')
+        "
 
     - name: Run tests
       working-directory: ./hindsight-api
@@ -146,6 +168,8 @@ jobs:
       HINDSIGHT_API_LLM_MODEL: openai/gpt-oss-20b
       HINDSIGHT_API_URL: http://localhost:8888
       GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+      # Prefer CPU-only PyTorch in CI (but keep PyPI for everything else)
+      UV_INDEX: pytorch=https://download.pytorch.org/whl/cpu
 
     steps:
     - uses: actions/checkout@v4
@@ -174,11 +198,11 @@ jobs:
 
     - name: Install client test dependencies
       working-directory: ./hindsight-clients/python
-      run: uv sync --extra test
+      run: uv sync --extra test --index-strategy unsafe-best-match
 
     - name: Install API dependencies
       working-directory: ./hindsight-api
-      run: uv sync
+      run: uv sync --no-install-project --index-strategy unsafe-best-match
 
     - name: Create .env file
       run: |
@@ -223,6 +247,8 @@ jobs:
       HINDSIGHT_API_LLM_MODEL: openai/gpt-oss-20b
       HINDSIGHT_API_URL: http://localhost:8888
       GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+      # Prefer CPU-only PyTorch in CI (but keep PyPI for everything else)
+      UV_INDEX: pytorch=https://download.pytorch.org/whl/cpu
 
     steps:
     - uses: actions/checkout@v4
@@ -252,7 +278,7 @@ jobs:
 
     - name: Install API dependencies
       working-directory: ./hindsight-api
-      run: uv sync
+      run: uv sync --no-install-project --index-strategy unsafe-best-match
 
     - name: Install TypeScript client dependencies
       working-directory: ./hindsight-clients/typescript
@@ -305,6 +331,8 @@ jobs:
       HINDSIGHT_API_LLM_MODEL: openai/gpt-oss-20b
       HINDSIGHT_API_URL: http://localhost:8888
       GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+      # Prefer CPU-only PyTorch in CI (but keep PyPI for everything else)
+      UV_INDEX: pytorch=https://download.pytorch.org/whl/cpu
 
     steps:
     - uses: actions/checkout@v4
@@ -341,7 +369,7 @@ jobs:
 
     - name: Install API dependencies
       working-directory: ./hindsight-api
-      run: uv sync
+      run: uv sync --no-install-project --index-strategy unsafe-best-match
 
     - name: Create .env file
       run: |

docker/standalone/Dockerfile

@@ -147,9 +147,13 @@ RUN mkdir -p /home/hindsight/.hindsight/bin && \
         curl -fsSL -o /home/hindsight/.hindsight/bin/pg0 \
             "https://github.com/vectorize-io/pg0/releases/latest/download/$PG0_BINARY" && \
         chmod +x /home/hindsight/.hindsight/bin/pg0 && \
+        ls -lh /home/hindsight/.hindsight/bin/pg0 && \
+        file /home/hindsight/.hindsight/bin/pg0 && \
+        ldd /home/hindsight/.hindsight/bin/pg0 2>&1 || true && \
         break || (echo "Retry $i failed, waiting..." && sleep 10); \
     done && \
-    /home/hindsight/.hindsight/bin/pg0 --version
+    echo "Testing pg0 binary..." && \
+    /home/hindsight/.hindsight/bin/pg0 --version || (echo "pg0 --version failed with exit code $?"; ldd /home/hindsight/.hindsight/bin/pg0; exit 1)
 
 # Pre-download PostgreSQL binaries
 ENV PG0_HOME=/home/hindsight/.pg0-cache
@@ -280,9 +284,13 @@ RUN mkdir -p /home/hindsight/.hindsight/bin && \
         curl -fsSL -o /home/hindsight/.hindsight/bin/pg0 \
             "https://github.com/vectorize-io/pg0/releases/latest/download/$PG0_BINARY" && \
         chmod +x /home/hindsight/.hindsight/bin/pg0 && \
+        ls -lh /home/hindsight/.hindsight/bin/pg0 && \
+        file /home/hindsight/.hindsight/bin/pg0 && \
+        ldd /home/hindsight/.hindsight/bin/pg0 2>&1 || true && \
         break || (echo "Retry $i failed, waiting..." && sleep 10); \
     done && \
-    /home/hindsight/.hindsight/bin/pg0 --version
+    echo "Testing pg0 binary..." && \
+    /home/hindsight/.hindsight/bin/pg0 --version || (echo "pg0 --version failed with exit code $?"; ldd /home/hindsight/.hindsight/bin/pg0; exit 1)
 
 # Pre-download PostgreSQL binaries
 ENV PG0_HOME=/home/hindsight/.pg0-cache

hindsight-api/hindsight_api/engine/cross_encoder.py

@@ -101,12 +101,7 @@ class LocalSTCrossEncoder(CrossEncoderModel):
             )
 
         logger.info(f"Reranker: initializing local provider with model {self.model_name}")
-        # Disable lazy loading (meta tensors) which causes issues with newer transformers/accelerate
+        self._model = CrossEncoder(self.model_name)
-        # Setting low_cpu_mem_usage=False and device_map=None ensures tensors are fully materialized
-        self._model = CrossEncoder(
-            self.model_name,
-            model_kwargs={"low_cpu_mem_usage": False, "device_map": None},
-        )
         logger.info("Reranker: local provider initialized")
 
     def predict(self, pairs: List[Tuple[str, str]]) -> List[float]:

hindsight-api/pyproject.toml

@@ -14,7 +14,7 @@ dependencies = [
     "openai>=1.0.0",
     "pydantic>=2.0.0",
     "rich>=13.0.0",
-    "sentence-transformers>=3.0.0",
+    "sentence-transformers>=3.0.0,<3.3.0",
     "langchain-text-splitters>=0.3.0",
     "fastapi[standard]>=0.120.3",
     "uvicorn>=0.38.0",
@@ -24,8 +24,8 @@ dependencies = [
     "pgvector>=0.4.1",
     "greenlet>=3.2.4",
     "psycopg2-binary>=2.9.11",
-    "transformers>=4.30.0",
+    "transformers>=4.30.0,<4.46.0",
-    "torch>=2.0.0",
+    "torch>=2.0.0,<2.6.0",
     "tiktoken>=0.12.0",
     "httpx>=0.27.0",
     "fastmcp>=2.0.0",

hindsight-api/tests/test_http_api_integration.py

@@ -281,7 +281,8 @@ async def test_full_api_workflow(api_client, test_bank_id):
     final_banks_data = response.json()["banks"]
     final_banks = [a["bank_id"] for a in final_banks_data]
     assert test_bank_id in final_banks
-    assert len(final_banks) >= len(initial_banks) + 1
+    # Don't assert count increases due to parallel test cleanup races
+    # Just verify our bank exists in the list
 
     # ================================================================
     # 10. Clean Up

hindsight-clients/python/tests/test_main_operations.py

@@ -239,7 +239,7 @@ class TestEndToEndWorkflow:
             bank_id=workflow_bank_id,
             query="What programming technologies do I use?",
         )
-        assert len(search_results) > 0
+        assert len(search_results.results) > 0
 
         # 4. Generate contextual answer
         reflect_response = client.reflect(