From 6d666f8b8b0603e1cec464f2987bc1470392f74c Mon Sep 17 00:00:00 2001
From: Oumaima Hamza <ohamza@oicr.on.ca>
Date: Mon, 1 Apr 2024 22:17:04 -0400
Subject: [PATCH 1/2] Set up Github Actions to update gene list

---
 .../workflows/updating_oncokb_database.yml    | 52 +++++++++++++++++++
 1 file changed, 52 insertions(+)
 create mode 100644 .github/workflows/updating_oncokb_database.yml

diff --git a/.github/workflows/updating_oncokb_database.yml b/.github/workflows/updating_oncokb_database.yml
new file mode 100644
index 000000000..059898b54
--- /dev/null
+++ b/.github/workflows/updating_oncokb_database.yml
@@ -0,0 +1,52 @@
+name: update oncokb curated gene List
+
+on:
+  schedule:
+    # Runs at 00:00 on the 1st day of every second month
+    - cron: '0 0 1 */2 *'
+  workflow_dispatch: {}
+    #run manually
+
+jobs:
+  update-gene-list:
+    runs-on: ubuntu-latest
+
+    steps:
+    - name: checkout repo
+      uses: actions/checkout@v4
+      with:
+        ref: release-1.6.0
+
+    - name: set up python
+      uses: actions/setup-python@v5
+      with:
+        python-version: '3.10'
+
+    - name: installing dependencies
+      run:
+        pip install pandas
+
+    - name: fetch and update gene list
+      env:
+        TOKEN: ${{ secrets.ONCOKB_TOKEN }}
+      run: |
+        # download the database
+        curl -H "Authorization: Bearer $TOKEN" https://www.oncokb.org/api/v1/utils/allCuratedGenes > allCuratedGenes.json
+        
+        # convert json to tsv
+        python -c "import pandas as pd; import datetime; df = pd.read_json('allCuratedGenes.json'); date_str = datetime.datetime.now().strftime('%Y%m%d'); df.to_csv(f'djerba/src/lib/djerba/data/{date_str}_allcuratedgenelist.tsv', sep='\t', index=False)"
+        
+        # delete json file 
+        rm allCuratedGenes.json
+        
+        # replace filename in the constants.py script
+        sed -i "s|ALL_CURATED_GENES = .*|ALL_CURATED_GENES = '{date_str}_allcuratedgenelist.tsv'|g" djerba/src/lib/djerba/util/oncoKB/constants.py
+
+    - name: create pull request
+      uses: peter-evans/create-pull-request@v6
+      with:
+        commit-message: "Update curated gene list"
+        branch: OncoKb_database_update
+        base: release-1.6.0
+        title: "Update curated gene list"
+        body: "This is an automated PR made by Github Actions to update OncoKb list of curated genes."

From c3934e5471592640aebe300a23a3735685cb1ca3 Mon Sep 17 00:00:00 2001
From: Oumaima Hamza <ohamza@oicr.on.ca>
Date: Tue, 2 Apr 2024 23:24:00 -0400
Subject: [PATCH 2/2] using awk to replace filename

---
 .github/workflows/updating_oncokb_database.yml | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/.github/workflows/updating_oncokb_database.yml b/.github/workflows/updating_oncokb_database.yml
index 059898b54..dbc289899 100644
--- a/.github/workflows/updating_oncokb_database.yml
+++ b/.github/workflows/updating_oncokb_database.yml
@@ -34,13 +34,13 @@ jobs:
         curl -H "Authorization: Bearer $TOKEN" https://www.oncokb.org/api/v1/utils/allCuratedGenes > allCuratedGenes.json
         
         # convert json to tsv
-        python -c "import pandas as pd; import datetime; df = pd.read_json('allCuratedGenes.json'); date_str = datetime.datetime.now().strftime('%Y%m%d'); df.to_csv(f'djerba/src/lib/djerba/data/{date_str}_allcuratedgenelist.tsv', sep='\t', index=False)"
+        python -c "import pandas as pd; import datetime; df = pd.read_json('allCuratedGenes.json'); date_str = datetime.datetime.now().strftime('%Y%m%d'); df.to_csv(f'src/lib/djerba/data/{date_str}_allcuratedgenelist.tsv', sep='\t', index=False)"
         
         # delete json file 
         rm allCuratedGenes.json
         
         # replace filename in the constants.py script
-        sed -i "s|ALL_CURATED_GENES = .*|ALL_CURATED_GENES = '{date_str}_allcuratedgenelist.tsv'|g" djerba/src/lib/djerba/util/oncoKB/constants.py
+        awk -v new_val="ALL_CURATED_GENES = \"${date_str}_allcuratedgenelist.tsv\"" '{gsub(/ALL_CURATED_GENES = \047.*\047/, new_val)}1' src/lib/djerba/util/oncoKB/constants.py > temp && mv temp src/lib/djerba/util/oncoKB/constants.py
 
     - name: create pull request
       uses: peter-evans/create-pull-request@v6