Add Ollama-backed Claude launcher with kimi-k2.7-code:cloud model

Modelfile

@@ -0,0 +1,9 @@
+FROM kimi-k2.7-code:cloud
+
+SYSTEM """You are an expert data scientist and Python/Pandas assistant.
+You help users understand and write Pandas code for real-life data analysis tasks.
+When asked about code, provide clear, working examples with brief explanations.
+Focus on practical, idiomatic Pandas usage."""
+
+PARAMETER temperature 0.2
+PARAMETER num_ctx 16384

claude

@@ -0,0 +1,188 @@
+#!/usr/bin/env python3
+"""
+Ollama-backed Claude-like assistant launcher.
+
+Usage:
+    ollama launch claude --model kimi-k2.7-code:cloud
+    ./claude --model kimi-k2.7-code:cloud
+"""
+
+import sys
+import argparse
+import json
+import os
+import urllib.request
+import urllib.error
+
+OLLAMA_BASE_URL = os.environ.get("OLLAMA_HOST", "http://localhost:11434")
+
+SYSTEM_PROMPT = """You are an expert data scientist and Python/Pandas assistant.
+You help users understand and write Pandas code for real-life data analysis tasks.
+When asked about code, provide clear, working examples with brief explanations.
+Focus on practical, idiomatic Pandas usage."""
+
+
+def check_ollama_running():
+    try:
+        req = urllib.request.urlopen(f"{OLLAMA_BASE_URL}/api/tags", timeout=3)
+        return req.status == 200
+    except Exception:
+        return False
+
+
+def list_local_models():
+    try:
+        req = urllib.request.urlopen(f"{OLLAMA_BASE_URL}/api/tags", timeout=5)
+        data = json.loads(req.read())
+        return [m["name"] for m in data.get("models", [])]
+    except Exception:
+        return []
+
+
+def stream_chat(model, messages):
+    payload = json.dumps({
+        "model": model,
+        "messages": messages,
+        "stream": True,
+    }).encode()
+
+    req = urllib.request.Request(
+        f"{OLLAMA_BASE_URL}/api/chat",
+        data=payload,
+        headers={"Content-Type": "application/json"},
+        method="POST",
+    )
+
+    full_response = ""
+    try:
+        with urllib.request.urlopen(req) as resp:
+            for line in resp:
+                line = line.strip()
+                if not line:
+                    continue
+                chunk = json.loads(line)
+                token = chunk.get("message", {}).get("content", "")
+                if token:
+                    print(token, end="", flush=True)
+                    full_response += token
+                if chunk.get("done"):
+                    break
+    except urllib.error.URLError as e:
+        print(f"\nError communicating with Ollama: {e}", file=sys.stderr)
+        sys.exit(1)
+
+    print()
+    return full_response
+
+
+def pull_model(model):
+    print(f"Pulling model '{model}'... (this may take a while)")
+    payload = json.dumps({"name": model, "stream": True}).encode()
+    req = urllib.request.Request(
+        f"{OLLAMA_BASE_URL}/api/pull",
+        data=payload,
+        headers={"Content-Type": "application/json"},
+        method="POST",
+    )
+    try:
+        with urllib.request.urlopen(req) as resp:
+            for line in resp:
+                line = line.strip()
+                if not line:
+                    continue
+                chunk = json.loads(line)
+                status = chunk.get("status", "")
+                if status:
+                    print(f"\r{status}", end="", flush=True)
+                if chunk.get("status") == "success":
+                    break
+        print("\nModel ready.")
+    except urllib.error.URLError as e:
+        print(f"\nFailed to pull model: {e}", file=sys.stderr)
+        sys.exit(1)
+
+
+def repl(model):
+    print(f"Claude (via Ollama / {model})")
+    print("Type your message and press Enter. Use Ctrl+C or 'exit' to quit.\n")
+
+    messages = [{"role": "system", "content": SYSTEM_PROMPT}]
+
+    while True:
+        try:
+            user_input = input("You: ").strip()
+        except (KeyboardInterrupt, EOFError):
+            print("\nGoodbye.")
+            break
+
+        if not user_input:
+            continue
+        if user_input.lower() in ("exit", "quit", "/exit", "/quit"):
+            print("Goodbye.")
+            break
+
+        messages.append({"role": "user", "content": user_input})
+        print("Assistant: ", end="", flush=True)
+        response = stream_chat(model, messages)
+        messages.append({"role": "assistant", "content": response})
+
+
+def main():
+    parser = argparse.ArgumentParser(
+        prog="claude",
+        description="Launch a Claude-like assistant backed by a local Ollama model.",
+    )
+    parser.add_argument(
+        "--model",
+        default="kimi-k2.7-code:cloud",
+        help="Ollama model to use (default: kimi-k2.7-code:cloud)",
+    )
+    parser.add_argument(
+        "--list-models",
+        action="store_true",
+        help="List locally available Ollama models and exit.",
+    )
+    parser.add_argument(
+        "--pull",
+        action="store_true",
+        help="Pull the model before starting (download if not present).",
+    )
+    args = parser.parse_args()
+
+    if not check_ollama_running():
+        print(
+            "Ollama is not running or not installed.\n"
+            "Install it from https://ollama.com and run: ollama serve\n"
+            "Then re-run this command.",
+            file=sys.stderr,
+        )
+        sys.exit(1)
+
+    if args.list_models:
+        models = list_local_models()
+        if models:
+            print("Available models:")
+            for m in models:
+                print(f"  {m}")
+        else:
+            print("No models found locally. Run with --pull to download one.")
+        return
+
+    local_models = list_local_models()
+    model = args.model
+
+    if model not in local_models:
+        if args.pull:
+            pull_model(model)
+        else:
+            print(
+                f"Model '{model}' is not available locally.\n"
+                f"Run with --pull to download it, or choose from: {local_models or ['(none)']}"
+            )
+            sys.exit(1)
+
+    repl(model)
+
+
+if __name__ == "__main__":
+    main()