use glob pattern for docs (#134)

* update, remove using library for mkdocs, use glob properly, do not run deploy on PR

Signed-off-by: Filinto Duran <[email protected]>

* simplify if

Signed-off-by: Filinto Duran <[email protected]>

---------

Signed-off-by: Filinto Duran <[email protected]>

Author: filintod

.github/workflows/docs.yml

@@ -4,12 +4,12 @@ on:
     branches:
       - main
     paths:
-      - docs
+      - docs/**
   pull_request:
     branches:
       - main
     paths:
-      - docs
+      - docs/**
   workflow_dispatch:
 permissions:
   contents: write
@@ -42,18 +42,24 @@ jobs:
       - name: Remove plugins from mkdocs configuration
         run: |
           sed -i '/^plugins:/,/^[^ ]/d' mkdocs.yml
-      - name: Run MkDocs build
-        uses: Kjuly/mkdocs-page-builder@main
+      - name: Install Python dependencies
+        run: |
+          pip install mkdocs-material
+          pip install .[recommended,git,imaging]
+          pip install mkdocs-jupyter
+      - name: Validate build
+        run: mkdocs build
 
   deploy:
+    if: github.ref == 'refs/heads/main'
     runs-on: ubuntu-latest
     needs: documentation_validation
     steps:
       - uses: actions/checkout@v4
       - uses: actions/setup-python@v5
         with:
           python-version: 3.x
-      - run: echo "cache_id=$(date --utc '+%V')" >> $GITHUB_ENV 
+      - run: echo "cache_id=$(date --utc '+%V')" >> $GITHUB_ENV
       - uses: actions/cache@v4
         with:
           key: mkdocs-material-${{ env.cache_id }}

docs/concepts/agents.md

@@ -121,4 +121,4 @@ The LLM-Based Workflow relies on the reasoning capabilities of an LLM to dynamic
 * Chat History: Previous agent responses and interactions.
 * Agent Metadata: Attributes like expertise, availability, or priorities.
 
-This approach ensures that the most suitable agent is selected for each task, optimizing collaboration and efficiency. For example, in a multi-agent customer support system, the LLM can assign tasks to agents based on customer issues, agent expertise, and workload distribution.
+This approach ensures that the most suitable agent is selected for each task, optimizing collaboration and efficiency. For example, in a multi-agent customer support system, the LLM can assign tasks to agents based on customer issues, agent expertise, and workload distribution.