readme

.gitignore

@@ -1,2 +1,3 @@
 */.cache
-application/flagged
+application/flagged
+qdrant_storage/

README.md

@@ -0,0 +1,24 @@
+# qdurllm
+
+**qdurllm** is a local search engine that lets you search through different URLs, retrieve and chat with their content, powered by Qdrant, Langchain, llama.cpp and Gradio.
+
+## Quickstart
+
+### Installation
+
+```bash
+git clone https://github.com/AstraBert/qdurllm.git
+cd qdurllm
+```
+
+### Run it!
+
+```bash
+docker compose up -d
+```
+
+You'll find the application running at `http://localhost:7860` or `http://0.0.0.0:7860`
+
+## Flowchart
+
+![Flowchart](imgs/qdurllm.png)

application/Dockerfile

@@ -1,12 +1,10 @@
-FROM python:3.10.14-slim-bookworm
+FROM astrabert/local-search-application:latest
 
 WORKDIR /app
 
-ADD . /app/
+RUN mkdir -p /app/scripts
 
-RUN python3 -m pip cache purge
-RUN python3 -m pip install --no-cache-dir -r requirements.txt
-RUN python3 -m nltk.downloader all
+ADD ./scripts/*.py /app/scripts/
 
 EXPOSE 7860
 

application/build_command.sh

@@ -1,5 +1,5 @@
 docker buildx build \
---label org.opencontainers.image.title=astrabert/local-search-application\
+--label org.opencontainers.image.title=astrabert/local-search-application \
 --label org.opencontainers.image.description='Local Search engine container' \
 --label org.opencontainers.image.url=https://github.com/AstraBert/qdurllm \
 --label org.opencontainers.image.source=https://github.com/AstraBert/qdurllm \

application/scripts/app.py

@@ -50,7 +50,7 @@ def reranked_rag(client, encoder0, encoder1, collection0, collection1, message):
 demo0 = gr.Interface(fn=call_upload2qdrant, title="Upload URL content to Qdrant", inputs=gr.Textbox(label="URL(s)", info="Add one URL or more (if more, you should provide them comma-separated, like this: URL1,URL2,...,URLn)"), outputs=gr.Textbox(label="Logs"))
 
 def llama_cpp_respond(query, max_new_tokens, temperature, repeat_penalty, seed):
-    url = "http://localhost:8000/completion"
+    url = "http://host.docker.internal:8000/completion"
     headers = {
         "Content-Type": "application/json"
     }
@@ -65,6 +65,7 @@ def llama_cpp_respond(query, max_new_tokens, temperature, repeat_penalty, seed):
     response = requests.post(url, headers=headers, json=data)
 
     a = response.json()
+    print(a)
     return a["content"]
 
 

compose.yaml

@@ -22,5 +22,5 @@ services:
     ports:
       - "8000:8000"
     networks:
-      - mynets
+      - mynet