ADD: 1. Initialize the repository and add basic info for GitHub projects with over 1k stars. 2. Add basic semantic search functionality for the GitHub repository.

chat.py

@@ -0,0 +1,78 @@
+from langchain_openai import OpenAIEmbeddings
+from langchain_community.vectorstores import FAISS
+
+import os
+import gradio as gr
+import json
+import os
+
+# get the root path of the project
+current_file_path = os.path.dirname(os.path.abspath(__file__))
+root_path = os.path.abspath(current_file_path)
+
+class RepoSearch:
+    def __init__(self):
+
+        db_path = os.path.join(root_path, "database", "faiss_index")
+        embeddings = OpenAIEmbeddings(api_key="sk-Mo5K9m2hKXjV1DeGeBAIzXLZFxxiOTvSwUoemKmfMXdmE9Bs", 
+                              base_url="https://api.wlai.vip/v1",
+                              model="text-embedding-3-small")
+        
+        assert os.path.exists(db_path), f"Database not found: {db_path}"
+        self.vector_db = FAISS.load_local(db_path, embeddings, 
+                                          allow_dangerous_deserialization=True)
+        
+    def search(self, query, k=10):
+        '''
+            name + description + html_url + topics
+        '''
+        results = self.vector_db.similarity_search(query + " technology", k=k)
+
+        simple_str = ""
+        for i, doc in enumerate(results):
+            content = json.loads(doc.page_content)
+            if content["description"] is None:
+                content["description"] = ""
+            desc = content["description"] if len(content["description"]) < 300 else content["description"][:300] + "..."
+            simple_str += f"\t**{i+1}. {content['name']}** || **Description:** {desc} || **Url:** {content['html_url']} \n"
+
+        return simple_str
+
+
+def main():
+    search = RepoSearch()
+
+    def respond(
+        prompt: str,
+        history,
+    ):
+        if not history:
+            history = [{"role": "system", "content": "You are a friendly chatbot"}]
+        history.append({"role": "user", "content": prompt})
+
+        yield history
+
+        response = {"role": "assistant", "content": ""}
+        response["content"] = search.search(prompt)
+        yield history + [response]
+
+    with gr.Blocks() as demo:
+        gr.Markdown("## Semantic github search (基于语义的 github 仓库搜索) 🌐")
+        chatbot = gr.Chatbot(
+            label="Agent",
+            type="messages",
+            avatar_images=(
+                None,
+                "https://img1.baidu.com/it/u=2193901176,1740242983&fm=253&fmt=auto&app=138&f=JPEG?w=500&h=500",
+            ),
+            height="65vh"
+        )
+        prompt = gr.Textbox(max_lines=2, label="Chat Message")
+        prompt.submit(respond, [prompt, chatbot], [chatbot])
+        prompt.submit(lambda: "", None, [prompt])
+
+    demo.launch(share=True)
+
+
+if __name__ == "__main__":
+    main()