[stable31] fix: narrow down pip's package search and other ci fixes (#238)

Backport of PR #236

- fix: narrow down pip's package search
- chore: bump runner ubuntu to 24.04
- fix(ci): use the cpu version of torch to not run out of space
- fix(ci): fix langchain package versions
- fix(ci): use the cpu version of torch in static analysis

Author: kyteinsky

.github/workflows/integration-test.yml

@@ -52,7 +52,7 @@ jobs:
               - '.github/workflows/integration-test.yml'
 
   integration:
-    runs-on: ubuntu-22.04
+    runs-on: ubuntu-24.04
 
     needs: changes
     if: needs.changes.outputs.src != 'false'
@@ -179,6 +179,9 @@ jobs:
         run: |
           cd context_chat_backend
           pip install --upgrade pip setuptools wheel
+          # use the cpu version of torch to not run out of space
+          pip install torch torchvision --index-url https://download.pytorch.org/whl/cpu
+          sed -i '/torch(vision)?/d' requirements.txt
           pip install -r requirements.txt
           cp example.env .env
           echo "NEXTCLOUD_URL=http://localhost:8080" >> .env

.github/workflows/static-analysis.yml

@@ -72,6 +72,9 @@ jobs:
       - name: Install dependencies
         run: |
           pip install --upgrade pip setuptools wheel
+          # use the cpu version of torch to not run out of space
+          pip install torch torchvision --index-url https://download.pytorch.org/whl/cpu
+          sed -i '/torch(vision)?/d' requirements.txt
           pip install -r requirements_dev.txt
 
       - name: Static analysis with pyright

requirements.txt

@@ -1,38 +1,40 @@
-apscheduler
-chromadb
-cryptography
-ctransformers
-epub2txt
-fastapi
-httpx
-langchain
-langchain-community
-langchain-postgres
-langchain-unstructured
-llama_cpp_python
-msg-parser
+apscheduler>=3.11.1
+chromadb>=1.3.5
+cryptography>=46.0.3
+ctransformers>=0.2.27
+docx2txt>=0.9
+epub2txt>=0.1.6
+fastapi>=0.122.0
+httpx>=0.28.1
+# fix the langchain package versions to avoid long computation in pip dep resolution
+# langchain is stuck at <1.0.0 due to langhchain-unstructured not being updated yet
+langchain==0.3.27
+langchain-community==0.3.21
+langchain-postgres==0.0.16
+langchain-unstructured==0.1.6
+llama_cpp_python>=0.3.16
+msg-parser>=1.2.0
 # fix it to the version before niquests
 nc_py_api==0.20.2
-odfdo
-odfpy
-openpyxl
-pandas
-psutil
-pypdf
-docx2txt
-protobuf==3.20.0
-python-dotenv
-python-multipart
-python-pptx
-ruamel.yaml
-sentence-transformers
-setuptools
-sse_starlette
-starlette_context
-striprtf
-torch
-torchvision
-transformers
+odfdo>=3.17.6
+odfpy>=1.4.1
+openpyxl>=3.1.5
+pandas>=2.3.3
+psutil>=7.1.3
+pypdf>=6.4.0
+protobuf>=6.33.1
+python-dotenv>=1.2.1
+python-multipart>=0.0.20
+python-pptx>=1.0.2
+ruamel.yaml>=0.18.16
+sentence-transformers>=5.1.2
+setuptools>=80.9.0
+sse_starlette>=3.0.3
+starlette_context>=0.4.0
+striprtf>=0.0.29
+torch>=2.9.1
+torchvision>=0.24.1
+transformers>=4.57.3
 unstructured @ git+https://github.com/kyteinsky/unstructured@68394f64e8fcf800945327eb246369ab31d95eca
-wheel
-xlrd
+wheel>=0.45.1
+xlrd>=2.0.2