Merge pull request #150 from helicalAI/main

Bring back skipped tests, fix seed for reproducible results in scGPT, include updates in README
helicalAI · Dec 11, 2024 · 1551b01 · 1551b01
2 parents 488294f + f5a3a12
commit 1551b01
Show file tree

Hide file tree

Showing 34 changed files with 569 additions and 221 deletions.
diff --git a/.github/workflows/main.yml b/.github/workflows/main.yml
@@ -20,6 +20,8 @@ jobs:
       - name: Install dependencies
         run: |
             pip install -r requirements-dev.txt
+            # Currently, the self-hosted runner uses a pre-installed version of mamba-ssm (v2.2.2). Using the latest version (v2.2.4) breaks the pip installation.
+            # TODO fix the installation of helical[mamba-ssm], and add it in the GitHub actions (CI/CD pipeline).
             pip install .
             
       # First download before tests as they make use of the downloaded files 
@@ -29,41 +31,93 @@ jobs:
 
       - name: Execute unittests
         run: |
-          pytest --cov-report=html:html_cov --cov-branch --cov-report term --cov=helical ci/
+          CUDA_VISIBLE_DEVICES=-1 pytest --cov-report=html:html_cov --cov-branch --cov-report term --cov=helical ci/
       
       - name: Upload coverage report
         uses: actions/upload-artifact@v4
         with:
           name: coverage-report
           path: html_cov/
+
+  integration-tests:
+    needs: tests
+    runs-on: self-hosted
+    env:
+      CUDA_VISIBLE_DEVICES: 0
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v2
 
-      # Does not seem to work but would be nice to have
-      # - name: Pytest coverage comment
-      #   uses: MishaKav/pytest-coverage-comment@main
-      #   with:
-      #     pytest-coverage-path: ./pytest-coverage.txt
-      #     junitxml-path: ./pytest.xml
+      - name: setup python
+        uses: actions/setup-python@v5
+        with:
+          python-version: 3.11.8
 
+      # because jobs may not be run in the same order, we need to install the dependencies again
+      - name: Install helical
+        run: |
+            pip install .
+        
+      # Required to get the data
+      - name: Download all files
+        run: |
+          python ci/download_all.py
 
       - name: Execute Geneformer v1
         run: |
-          python examples/run_models/run_geneformer.py ++model_name="gf-12L-30M-i2048"
+          python examples/run_models/run_geneformer.py ++model_name="gf-12L-30M-i2048" ++device="cuda"
+
+      - name: Fine-tune Geneformer v1
+        run: |
+          python examples/fine_tune_models/fine_tune_geneformer.py ++model_name="gf-12L-30M-i2048" ++device="cuda"
 
       - name: Execute Geneformer v2
         run: |
-          python examples/run_models/run_geneformer.py ++model_name="gf-12L-95M-i4096"
+          python examples/run_models/run_geneformer.py ++model_name="gf-12L-95M-i4096" ++device="cuda"
+
+      - name: Fine-tune Geneformer v2
+        run: |
+          python examples/fine_tune_models/fine_tune_geneformer.py ++model_name="gf-12L-30M-i2048" ++device="cuda"
 
       - name: Execute scGPT
         run: |
-          python examples/run_models/run_scgpt.py
+          python examples/run_models/run_scgpt.py ++device="cuda"
+
+      - name: Fine-tune scGPT
+        run: |
+          python examples/fine_tune_models/fine_tune_scgpt.py ++device="cuda"
 
       - name: Execute UCE
         run: |
-          python examples/run_models/run_uce.py
+          python examples/run_models/run_uce.py ++device="cuda"
+
+      - name: Fine-tune UCE
+        run: |
+          python examples/fine_tune_models/fine_tune_UCE.py ++device="cuda"
+
+      - name: Execute Hyena
+        run: |
+          python examples/run_models/run_hyena_dna.py ++device="cuda"
+
+      - name: Execute Hyena
+        run: |
+          python examples/fine_tune_models/fine_tune_hyena_dna.py ++device="cuda"
+
+      - name: Execute Helix-mRNA
+        run: |
+          python examples/run_models/run_helix_mrna.py ++device="cuda"
 
-      # - name: Execute Hyena
-      #   run: |
-      #     python examples/run_models/run_hyena_dna.py
+      - name: Fine-tune Helix-mRNA
+        run: |
+          python examples/fine_tune_models/fine_tune_helix_mrna.py ++device="cuda"
+
+      - name: Execute Mamba2-mRNA
+        run: |
+          python examples/run_models/run_mamba2_mrna.py ++device="cuda"
+
+      - name: Fine-tune Mamba2-mRNA
+        run: |
+          python examples/fine_tune_models/fine_tune_mamba2_mrna.py ++device="cuda"
 
       - name: Execute benchmarking
         run: |
@@ -73,6 +127,8 @@ jobs:
   notebooks:
     needs: tests
     runs-on: self-hosted
+    env:
+      CUDA_VISIBLE_DEVICES: 0
     steps:
       - name: Checkout repository
         uses: actions/checkout@v2
@@ -82,10 +138,11 @@ jobs:
         with:
           python-version: 3.11.8
 
-      - name: Install dependencies
+      # because jobs may not be run in the same order, we need to install the dependencies again
+      - name: Install helical
         run: |
-            pip install -r requirements-dev.txt
-            
+            pip install .
+        
       - name: Reduce datasets to speedup checks
         run: |
           sed -i 's/train\[:65%\]/train\[:5%\]/g' ./examples/notebooks/Cell-Type-Annotation.ipynb
@@ -97,4 +154,4 @@ jobs:
 
       - name: Run Notebooks
         run: |
-          pytest --durations=0 --nbmake $(find ./examples/notebooks -name "*.ipynb" ! -name "Hyena*")
+          pytest --durations=0 --nbmake ./examples/notebooks/*.ipynb
diff --git a/.github/workflows/release.yml b/.github/workflows/release.yml
@@ -9,6 +9,21 @@ permissions:
   contents: write
 
 jobs:
+  fresh-install:
+    runs-on: ubuntu-20.04
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v2
+
+      - name: setup python
+        uses: actions/setup-python@v5
+        with:
+          python-version: 3.11.8
+
+      - name: Install dependencies
+        run: |
+            pip install .
+
   tests:
     runs-on: self-hosted
     steps:
@@ -23,6 +38,8 @@ jobs:
       - name: Install dependencies
         run: |
             pip install -r requirements-dev.txt
+            # Currently, the self-hosted runner uses a pre-installed version of mamba-ssm (v2.2.2). Using the latest version (v2.2.4) breaks the pip installation.
+            # TODO fix the installation of helical[mamba-ssm], and add it in the GitHub actions (CI/CD pipeline).
             pip install .
             
       # First download before tests as they make use of the downloaded files 
@@ -32,41 +49,93 @@ jobs:
 
       - name: Execute unittests
         run: |
-          pytest --cov-report=html:html_cov --cov-branch --cov-report term --cov=helical ci/
+          CUDA_VISIBLE_DEVICES=-1 pytest --cov-report=html:html_cov --cov-branch --cov-report term --cov=helical ci/
       
       - name: Upload coverage report
         uses: actions/upload-artifact@v4
         with:
           name: coverage-report
           path: html_cov/
 
-      # Does not seem to work but would be nice to have
-      # - name: Pytest coverage comment
-      #   uses: MishaKav/pytest-coverage-comment@main
-      #   with:
-      #     pytest-coverage-path: ./pytest-coverage.txt
-      #     junitxml-path: ./pytest.xml
+  integration-tests:
+    needs: tests
+    runs-on: self-hosted
+    env:
+      CUDA_VISIBLE_DEVICES: 0
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v2
 
+      - name: setup python
+        uses: actions/setup-python@v5
+        with:
+          python-version: 3.11.8
+
+      # because jobs may not be run in the same order, we need to install the dependencies again
+      - name: Install helical
+        run: |
+            pip install .
+        
+      # Required to get the data
+      - name: Download all files
+        run: |
+          python ci/download_all.py
 
       - name: Execute Geneformer v1
         run: |
-          python examples/run_models/run_geneformer.py ++model_name="gf-12L-30M-i2048"
+          python examples/run_models/run_geneformer.py ++model_name="gf-12L-30M-i2048" ++device="cuda"
+
+      - name: Fine-tune Geneformer v1
+        run: |
+          python examples/fine_tune_models/fine_tune_geneformer.py ++model_name="gf-12L-30M-i2048" ++device="cuda"
 
       - name: Execute Geneformer v2
         run: |
-          python examples/run_models/run_geneformer.py ++model_name="gf-12L-95M-i4096"
+          python examples/run_models/run_geneformer.py ++model_name="gf-12L-95M-i4096" ++device="cuda"
+
+      - name: Fine-tune Geneformer v2
+        run: |
+          python examples/fine_tune_models/fine_tune_geneformer.py ++model_name="gf-12L-30M-i2048" ++device="cuda"
 
       - name: Execute scGPT
         run: |
-          python examples/run_models/run_scgpt.py
+          python examples/run_models/run_scgpt.py ++device="cuda"
+
+      - name: Fine-tune scGPT
+        run: |
+          python examples/fine_tune_models/fine_tune_scgpt.py ++device="cuda"
 
       - name: Execute UCE
         run: |
-          python examples/run_models/run_uce.py
+          python examples/run_models/run_uce.py ++device="cuda"
+
+      - name: Fine-tune UCE
+        run: |
+          python examples/fine_tune_models/fine_tune_UCE.py ++device="cuda"
+
+      - name: Execute Hyena
+        run: |
+          python examples/run_models/run_hyena_dna.py ++device="cuda"
 
-      # - name: Execute Hyena
-      #   run: |
-      #     python examples/run_models/run_hyena_dna.py
+      - name: Fine-tune Hyena
+        run: |
+          python examples/fine_tune_models/fine_tune_hyena_dna.py ++device="cuda"
+
+      - name: Execute Helix-mRNA
+        run: |
+          python examples/run_models/run_helix_mrna.py ++device="cuda"
+
+      - name: Fine-tune Helix-mRNA
+        run: |
+          python examples/fine_tune_models/fine_tune_helix_mrna.py ++device="cuda"
+
+      - name: Execute Mamba2-mRNA
+        run: |
+          python examples/run_models/run_mamba2_mrna.py ++device="cuda"
+
+      - name: Fine-tune Mamba2-mRNA
+        run: |
+          python examples/fine_tune_models/fine_tune_mamba2_mrna.py ++device="cuda"
 
       - name: Execute benchmarking
         run: |
@@ -76,6 +145,8 @@ jobs:
   notebooks:
     needs: tests
     runs-on: self-hosted
+    env:
+      CUDA_VISIBLE_DEVICES: 0
     steps:
       - name: Checkout repository
         uses: actions/checkout@v2
@@ -85,10 +156,11 @@ jobs:
         with:
           python-version: 3.11.8
 
-      - name: Install dependencies
+      # because jobs may not be run in the same order, we need to install the dependencies again
+      - name: Install helical
         run: |
-            pip install -r requirements-dev.txt
-            
+            pip install .
+                    
       - name: Reduce datasets to speedup checks
         run: |
           sed -i 's/train\[:65%\]/train\[:5%\]/g' ./examples/notebooks/Cell-Type-Annotation.ipynb
@@ -100,7 +172,7 @@ jobs:
 
       - name: Run Notebooks
         run: |
-          pytest --durations=0 --nbmake $(find ./examples/notebooks -name "*.ipynb" ! -name "Hyena*")
+          pytest --durations=0 --nbmake ./examples/notebooks/*.ipynb
 
   release:
     needs: notebooks
@@ -121,8 +193,8 @@ jobs:
       - name: Extract version
         id: get_version
         run: |
-          # Extract version from setup.py (or adjust for your version file)
-          VERSION=$(python setup.py --version)
+          # Extract version from pyproject.toml (or adjust for your version file)
+          VERSION=$(grep "version =" pyproject.toml | cut -d '"' -f 2)
           echo "::set-output name=version::$VERSION"
         
       - name: Check if tag exists

diff --git a/.readthedocs.yaml b/.readthedocs.yaml
@@ -8,9 +8,6 @@ build:
     pre_build:
       - "mkdir -p ./docs/notebooks"
       - "cp ./examples/notebooks/*.ipynb ./docs/notebooks"
-      - "pip install -r ./docs/requirements.txt"
-      # Build the MkDocs site
-      # - "mike deploy --push --update-aliases 1.0 latest"
 
 mkdocs:
   configuration: mkdocs.yml

diff --git a/CITATION.bib b/CITATION.bib
@@ -1,10 +1,10 @@
 @software{allard_2024_13135902,
   author       = {Helical Team},
-  title        = {helicalAI/helical: v0.0.1-alpha8},
+  title        = {helicalAI/helical: v0.0.1a14},
   month        = nov,
   year         = 2024,
   publisher    = {Zenodo},
-  version      = {0.0.1a6},
+  version      = {0.0.1a14},
   doi          = {10.5281/zenodo.13135902},
   url          = {https://doi.org/10.5281/zenodo.13135902}
 }