Initial commit

(cherry picked from commit 180c34177b38d123d03b31577048d242fb784ba0)

.gitignore

@@ -0,0 +1,178 @@
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#  Usually these files are written by a python script from a template
+#  before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py,cover
+.hypothesis/
+.pytest_cache/
+cover/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+.pybuilder/
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+#   For a library or package, you might want to ignore these files since the code is
+#   intended to run in multiple environments; otherwise, check them in:
+# .python-version
+
+# pipenv
+#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
+#   However, in case of collaboration, if having platform-specific dependencies or dependencies
+#   having no cross-platform support, pipenv may install dependencies that don't work, or not
+#   install all needed dependencies.
+#Pipfile.lock
+
+# UV
+#   Similar to Pipfile.lock, it is generally recommended to include uv.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#uv.lock
+
+# poetry
+#   Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#   https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
+#poetry.lock
+
+# pdm
+#   Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
+#pdm.lock
+#   pdm stores project-wide configurations in .pdm.toml, but it is recommended to not include it
+#   in version control.
+#   https://pdm.fming.dev/latest/usage/project/#working-with-version-control
+.pdm.toml
+.pdm-python
+.pdm-build/
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/
+
+# pytype static type analyzer
+.pytype/
+
+# Cython debug symbols
+cython_debug/
+
+# PyCharm
+#  JetBrains specific template is maintained in a separate JetBrains.gitignore that can
+#  be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
+#  and can be added to the global gitignore or merged into this file.  For a more nuclear
+#  option (not recommended) you can uncomment the following to ignore the entire idea folder.
+.idea/
+
+# Ruff stuff:
+.ruff_cache/
+
+# PyPI configuration file
+.pypirc
+
+# Qdrant storage
+.scripts/qdrant_storage
+

.scripts/run-qdrant.sh

@@ -0,0 +1,25 @@
+# Get the directory of this script
+script_dir="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+parent_dir="$(dirname "$script_dir")"
+data_dir="$parent_dir/.data"
+snapshot_file="$(cd "$data_dir" && pwd)/mcp-server-qdrant-knowledge-base.snapshot"
+
+# Start the container
+docker run -d \
+    --name "qdrant" \
+    -p 6333:6333 -p 6334:6334 \
+    -v "$(pwd)/qdrant_storage:/qdrant/storage:z" \
+    "qdrant/qdrant:v1.13.4"
+
+# Wait for Qdrant to start
+until $(curl --output /dev/null --silent --fail http://localhost:6333/healthz); do
+    printf '.'
+    sleep 5
+done
+
+# Load the data snapshot into Qdrant
+curl -X POST \
+    -H "Content-Type: multipart/form-data" \
+    -F "snapshot=@$snapshot_file" \
+    http://localhost:6333/collections/mcp-server-qdrant-knowledge-base/snapshots/upload?priority=snapshot
+

README.md

@@ -9,4 +9,121 @@ license: apache-2.0
 short_description: YouTube In-Video Search
 ---
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+# webinar-vibe-coding-rag
+
+**This repository contains materials for the hands-on "[Letting LLMs Write RAG 
+Applications](https://try.qdrant.tech/llm-rag)" webinar.**
+
+## Project Overview: YouTube In-Video Search
+
+When learning a new skill, YouTube videos can be a great resource. However, in-depth content is often lengthy and may 
+assume no prior knowledge. What if you could have a smart assistant to help you navigate through videos and find exactly 
+what you need? This project creates a search engine for video content, helping you skim through and focus on what 
+matters specifically to you.
+
+Retrieval Augmented Generation (RAG) is perfect for this task. By indexing a video's transcript, we provide an interface 
+to search through its content. Users can click on search results to jump to the exact timestamp where a topic is 
+discussed.
+
+### How It Works
+
+The application has two main views:
+
+1. **Input View**: 
+   - User provides a YouTube video URL
+   - Backend processes the video by:
+     - Extracting the transcript
+     - Dividing it into 30-second chunks with 10-second overlaps
+     - Creating embeddings using SentenceTransformers
+     - Storing these vectors in Qdrant
+
+2. **Video View**:
+   - Displays the video alongside its transcription
+   - Allows clicking on timestamps to jump to specific parts
+   - Provides a search bar to filter content
+   - When a query is submitted, backend returns the most relevant video segments
+   - Results appear as clickable links, while hiding irrelevant parts of the transcript
+
+If a video has been processed previously, the application uses existing embeddings without reprocessing.
+
+### Technologies Used
+
+This project uses the following tools:
+
+- [Qdrant](https://qdrant.tech/) - Vector search engine for both coding and in-video search
+- [SentenceTransformers](https://www.sbert.net/) - Pre-trained models for sentence embeddings (using 
+  `sentence-transformers/static-retrieval-mrl-en-v1`)
+- [FastAPI](https://fastapi.tiangolo.com/) - Framework for the REST API and serving the frontend
+- [DaisyUI](https://daisyui.com/) - Reusable frontend components for Tailwind CSS
+- Pure HTML, CSS, and vanilla JavaScript
+
+## Setup Instructions
+
+### Setting up mcp-server-qdrant
+
+The `mcp-server-qdrant` exposes two tools that interact with Qdrant:
+
+- `qdrant-find` - Search for similar entries in the Qdrant index
+- `qdrant-store` - Store new entries in the Qdrant index for future reference
+
+The LLM decides when to use these tools based on their descriptions.
+
+#### Configuring Tool Descriptions
+
+Configure the tools using environmental variables:
+
+```bash
+export TOOL_FIND_DESCRIPTION="Use this tool ALWAYS before generating any FRONTEND code. \
+It lets you search for relevant code snippets based on natural language descriptions. \
+The 'query' parameter should describe what you're looking for, and the tool will return the most relevant code \
+snippets. If this tool finds something similar, then create your code so it is consistent. Reuse existing code \
+as much as you can."
+
+export TOOL_STORE_DESCRIPTION="Store reusable FRONTEND code snippets for later retrieval. \
+The 'information' parameter should contain a natural language description of what the code does, while the actual \
+code should be included in the 'metadata' parameter as a 'code' property. The value of 'metadata' is a Python \
+dictionary with strings as keys. Use this always when you generate some code to store it for further reference."
+```
+
+**Note:** You can customize these descriptions to better suit your specific use case.
+
+#### Adding MCP Server to Claude Code
+
+Add the `mcp-server-qdrant` to available tools in Claude Code:
+
+```bash
+claude mcp add qdrant-code-search \
+  -e QDRANT_URL="http://localhost:6333" \
+  -e COLLECTION_NAME="mcp-server-qdrant-knowledge-base" \
+  -e TOOL_FIND_DESCRIPTION="$TOOL_FIND_DESCRIPTION" \
+  -e TOOL_STORE_DESCRIPTION="$TOOL_STORE_DESCRIPTION" \
+  -- uvx mcp-server-qdrant
+```
+
+Claude Code should always use the `qdrant-code-search` MCP before generating any code. When we accept generated code,
+it should store it in the `qdrant-code-search` server for future reference.
+
+### Initializing the Component Knowledge Base
+
+We'll use a specific version of [DaisyUI](https://daisyui.com/) that the LLM may not be trained on. Qdrant will serve as 
+a knowledge base for the LLM by storing DaisyUI components.
+
+Run the `.scripts/run-qdrant.sh` script to load DaisyUI components into Qdrant. You can view the collection's content in 
+the Web UI at [http://localhost:6333/dashboard](http://localhost:6333/dashboard).
+
+## Running the Application
+
+Now we can start the actual vibe coding. **Note: We cannot guarantee the LLM will generate code that works out of the 
+box, but let's try!**
+
+### Example YouTube Videos
+
+Andrej Karpathy publishes many interesting but lengthy videos on his YouTube channel. Here are some examples you can use 
+with this application:
+
+- [\[1hr Talk\] Intro to Large Language Models](https://www.youtube.com/watch?v=zjkBMFhNj_g)
+- [Deep Dive into LLMs like ChatGPT](https://www.youtube.com/watch?v=7xTGNNLPyMI)
+- [How I use LLMs](https://www.youtube.com/watch?v=EWvNQjAaOHw)
+
+If our vibe coding session is successful, we should be able to process these videos and search through their content 
+effectively.

poetry.lock

@@ -0,0 +1,7 @@
+# This file is automatically @generated by Poetry 1.8.3 and should not be changed by hand.
+package = []
+
+[metadata]
+lock-version = "2.0"
+python-versions = "^3.10"
+content-hash = "53f2eabc9c26446fbcc00d348c47878e118afc2054778c3c803a0a8028af27d9"

poetry.toml

@@ -0,0 +1,2 @@
+[virtualenvs]
+in-project = true

pyproject.toml

@@ -0,0 +1,22 @@
+[tool.poetry]
+name = "webinar-vibe-coding-rag"
+version = "0.1.0"
+description = "An attempt to live code a working Retrieval Augmented Generation app with AI coding tools"
+authors = ["Kacper Łukawski <[email protected]>"]
+readme = "README.md"
+package-mode = false
+
+[tool.poetry.dependencies]
+python = "^3.10"
+torch = {version = "^2.6.0+cpu", source = "pytorch-cpu"}
+sentence-transformers = "^3.4.1"
+qdrant-client = "^1.13.3"
+
+[[tool.poetry.source]]
+name = "pytorch-cpu"
+url = "https://download.pytorch.org/whl/cpu"
+priority = "explicit"
+
+[build-system]
+requires = ["poetry-core"]
+build-backend = "poetry.core.masonry.api"