deploy workflow file

.github/workflows/deploy.yml

@@ -6,5 +6,30 @@ on:
       - prod
 
 jobs:
-  deploy:
+  build-and-deploy:
+    name: run docker container on gc
+    runs-on: ubuntu-latest
+
     steps:
+      - name: Checkout Repository
+        uses: actions/checkout@v2
+
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v1
+
+      - name: Build Docker Container
+        run: |
+          docker buildx build --platform linux/amd64 -t chat-agh .
+
+      - name: Tag Docker Container
+        run: |
+          docker tag chat-agh europe-west4-docker.pkg.dev/chat-agh-417420/docker-repository/chat-agh
+
+      - name: Push Docker Container
+        run: |
+          docker push europe-west4-docker.pkg.dev/chat-agh-417420/docker-repository/chat-agh
+
+      - name: Terraform Apply
+        working-directory: infra
+        run: |
+          source .env && terraform apply --auto-approve

src/sources/urls_finder.py

@@ -39,15 +39,28 @@ def get_pages_from_sitemap(domain):
         'https://wh.agh.edu.pl',
         'https://www.sjo.agh.edu.pl',
         'https://www.swfis.agh.edu.pl',
-        'https://sylabusy.agh.edu.pl'
+        'https://sylabusy.agh.edu.pl',
+        'https://sylabusy.agh.edu.pl/pl/'
+        'https://skn.agh.edu.pl',
+        'https://dss.agh.edu.pl',
+        'https://akademik.agh.edu.pl'
     ]
 
     def format_link(link: str):
         return link.split('#')[0]
 
+    def filter(page):
+        for year in range(16, 22):
+            if f'20{str(year)}' in page:
+                return True
+        if len(page) > 110:
+            return True
+
+        return False
+
     pages = []
     for domain in domains:
-        new_links = [format_link(page) for page in get_pages_from_sitemap(domain)]
+        new_links = [format_link(page) for page in get_pages_from_sitemap(domain) if not filter(page)]
         pages.extend(new_links)
         print(f"{len(new_links)} url's found in domain {domain}")