drift-ml 0.1.10__tar.gz → 0.2.1__tar.gz

drift_ml-0.2.1/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2025 Lakshit Sachdeva
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

drift_ml-0.2.1/PKG-INFO

@@ -0,0 +1,10 @@
+Metadata-Version: 2.4
+Name: drift-ml
+Version: 0.2.1
+Summary: drift — terminal-first, chat-based AutoML. Open source. No tokens. No auth.
+Project-URL: Homepage, https://github.com/lakshitsachdeva/intent2model
+Project-URL: Repository, https://github.com/lakshitsachdeva/intent2model
+Requires-Python: >=3.10
+License-File: LICENSE
+Requires-Dist: requests>=2.28.0
+Dynamic: license-file

drift_ml-0.2.1/README.md

@@ -0,0 +1,79 @@
+# drift
+
+**Terminal-first, chat-based AutoML.** Open source. No tokens. No auth.
+
+---
+
+## Install
+
+```bash
+pipx install drift-ml
+```
+
+Or:
+
+```bash
+npm install -g drift-ml
+```
+
+(Both require `pipx install drift-ml` for the Python CLI.)
+
+---
+
+## Run
+
+```bash
+drift
+```
+
+On first run, drift downloads and starts the engine automatically. No backend setup. No config. No tokens.
+
+---
+
+## How it works
+
+- **Local-first** — Engine runs on your machine. Data never leaves.
+- **Chat-based** — `load data.csv`, `predict price`, `try something stronger`
+- **Auto-start** — Engine downloads and starts in the background. You never start a backend manually.
+- **No auth** — No API keys for drift. (You need an LLM for training: Ollama, Gemini CLI, etc.)
+
+---
+
+## Example
+
+```text
+drift › load iris.csv
+drift › predict variety
+drift › try something stronger
+drift › quit
+```
+
+---
+
+## Philosophy
+
+drift should feel like `git`, `docker`, `brew` — a tool you trust immediately. Zero friction. Open source.
+
+---
+
+## Web UI (optional)
+
+This repo also includes a web UI. For development:
+
+```bash
+./start.sh
+```
+
+Then open http://localhost:3000
+
+---
+
+## Contributing
+
+See [CONTRIBUTING.md](CONTRIBUTING.md).
+
+---
+
+## License
+
+MIT

{drift_ml-0.1.10 → drift_ml-0.2.1}/drift/engine_launcher.py

@@ -1,11 +1,10 @@
 """
 Download and start the drift engine binary when no backend is running.
-Makes pipx install drift-ml work standalone (no npm needed).
+Open source — no tokens, no auth. Downloads from public GitHub Releases.
 """
 
 import os
 import platform
-import shutil
 import subprocess
 import sys
 from pathlib import Path
@@ -16,9 +15,8 @@ try:
 except ImportError:
     requests = None
 
-ENGINE_TAG = "v0.1.4"
-GITHUB_REPO = "lakshitsachdeva/drift"
-GITHUB_API = f"https://api.github.com/repos/{GITHUB_REPO}/releases/tags/{ENGINE_TAG}"
+GITHUB_REPO = "lakshitsachdeva/drift"  # Engine binaries (v0.2.0+ also in intent2model)
+ENGINE_TAG = "v0.2.0"  # Pinned — direct URL, no API, no rate limits
 ENGINE_PORT = os.environ.get("DRIFT_ENGINE_PORT", "8000")
 HEALTH_URL = f"http://127.0.0.1:{ENGINE_PORT}/health"
 
@@ -60,78 +58,63 @@ def _engine_running() -> bool:
         return False
 
 
-def _get_asset_download_url(asset_name: str) -> str:
-    """Resolve GitHub release asset to download URL. Prefer direct URL (no API)."""
+def _get_asset_url(asset_name: str) -> str:
+    """Direct download URL. No API — avoids rate limits."""
     base = os.environ.get("DRIFT_ENGINE_BASE_URL")
     if base:
         return f"{base.rstrip('/')}/{asset_name}"
     return f"https://github.com/{GITHUB_REPO}/releases/download/{ENGINE_TAG}/{asset_name}"
 
 
-def _get_asset_download_url_via_api(asset_name: str) -> str:
-    """Fallback: get URL via GitHub API (for private repos)."""
-    token = os.environ.get("DRIFT_GITHUB_TOKEN") or os.environ.get("GITHUB_TOKEN")
-    headers = {
-        "User-Agent": "Drift-Engine-Launcher/1.0",
-        "Accept": "application/vnd.github+json",
-    }
-    if token:
-        headers["Authorization"] = f"Bearer {token}"
-    r = requests.get(GITHUB_API, headers=headers, timeout=15)
-    if r.status_code == 404:
-        raise RuntimeError(
-            f"Release {ENGINE_TAG} not found. "
-            "If the repo is private, set DRIFT_GITHUB_TOKEN with repo read access."
-        )
-    r.raise_for_status()
-    data = r.json()
-    for a in data.get("assets", []):
-        if a.get("name") == asset_name:
-            url = a.get("browser_download_url") or a.get("url")
-            if url:
-                return url
-    raise RuntimeError(f"Asset {asset_name} not found in release {ENGINE_TAG}")
-
-
 def _download_file(url: str, dest: Path) -> None:
     dest.parent.mkdir(parents=True, exist_ok=True)
-    token = os.environ.get("DRIFT_GITHUB_TOKEN") or os.environ.get("GITHUB_TOKEN")
-    headers = {"User-Agent": "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36"}
-    if "api.github.com" in url:
-        headers["Accept"] = "application/octet-stream"
-    if token:
-        headers["Authorization"] = f"Bearer {token}"
-    try:
+    headers = {"User-Agent": "Drift/1.0"}
+
+    def try_curl():
+        result = subprocess.run(
+            ["curl", "-fsSL", "-o", str(dest), url],
+            capture_output=True,
+            timeout=120,
+        )
+        if result.returncode != 0:
+            raise RuntimeError("curl failed")
+
+    def try_requests():
         r = requests.get(url, headers=headers, stream=True, timeout=60)
         r.raise_for_status()
         with open(dest, "wb") as f:
             for chunk in r.iter_content(chunk_size=65536):
                 f.write(chunk)
-    except Exception:
-        # Fallback: urllib (stdlib, no deps, works when requests has issues)
+
+    def try_urllib():
         import urllib.request
         req = urllib.request.Request(url, headers=headers)
         with urllib.request.urlopen(req, timeout=60) as resp:
             with open(dest, "wb") as f:
                 f.write(resp.read())
 
+    for attempt in [try_curl, try_requests, try_urllib]:
+        try:
+            attempt()
+            return
+        except Exception:
+            continue
+    raise RuntimeError(f"Download failed. Try: curl -L -o {dest} {url}")
+
 
 def ensure_engine() -> bool:
-    """
-    If engine not running: download (if needed), start it, wait for health.
-    Returns True if engine is ready, False on failure.
-    """
+    """Download (if needed), start engine, wait for health."""
     if _engine_running():
         return True
 
     if not requests:
-        print("drift: 'requests' required for engine download. pip install requests", file=sys.stderr)
+        print("drift: 'requests' required. pip install requests", file=sys.stderr)
         return False
 
     bin_path = _get_engine_path()
     bin_dir = _get_engine_dir()
     if not bin_path or not bin_dir:
-        print("drift: Could not resolve engine dir (~/.drift/bin). Set HOME or USERPROFILE.", file=sys.stderr)
+        print("drift: Could not resolve ~/.drift/bin. Set HOME or USERPROFILE.", file=sys.stderr)
         return False
 
     bin_dir.mkdir(parents=True, exist_ok=True)
@@ -141,17 +124,13 @@ def ensure_engine() -> bool:
         ext = ".exe" if platform.system() == "Windows" else ""
         asset = f"drift-engine-{plat}-{arch}{ext}"
         print(f"drift: Downloading engine ({asset})...", file=sys.stderr)
-        url = _get_asset_download_url(asset)
         try:
+            url = _get_asset_url(asset)
             _download_file(url, bin_path)
         except Exception as e:
-            try:
-                url = _get_asset_download_url_via_api(asset)
-                _download_file(url, bin_path)
-            except Exception as e2:
-                print(f"drift: Download failed: {e2}", file=sys.stderr)
-                print(f"drift: Try manually: curl -L -o {bin_path} {url}", file=sys.stderr)
-                return False
+            print(f"drift: Download failed: {e}", file=sys.stderr)
+            print(f"drift: Run: mkdir -p ~/.drift/bin && curl -L -o ~/.drift/bin/{asset} <url>", file=sys.stderr)
+            return False
         if platform.system() != "Windows":
             bin_path.chmod(0o755)
         if platform.system() == "Darwin":
@@ -160,7 +139,13 @@ def ensure_engine() -> bool:
             except Exception:
                 pass
 
-    # Start engine
+    if platform.system() == "Darwin" and bin_path:
+        try:
+            bin_path.chmod(0o755)
+            subprocess.run(["xattr", "-dr", "com.apple.quarantine", str(bin_path)], check=False, capture_output=True)
+        except Exception:
+            pass
+
     env = {**os.environ, "DRIFT_ENGINE_PORT": ENGINE_PORT}
     proc = subprocess.Popen(
         [str(bin_path)],
@@ -170,10 +155,9 @@ def ensure_engine() -> bool:
         stderr=subprocess.DEVNULL,
         start_new_session=True,
     )
-    proc.wait()  # wait briefly to avoid race
+    proc.wait()
     del proc
 
-    # Poll for health
     for _ in range(60):
         if _engine_running():
             return True

drift_ml-0.2.1/drift_ml.egg-info/PKG-INFO

@@ -0,0 +1,10 @@
+Metadata-Version: 2.4
+Name: drift-ml
+Version: 0.2.1
+Summary: drift — terminal-first, chat-based AutoML. Open source. No tokens. No auth.
+Project-URL: Homepage, https://github.com/lakshitsachdeva/intent2model
+Project-URL: Repository, https://github.com/lakshitsachdeva/intent2model
+Requires-Python: >=3.10
+License-File: LICENSE
+Requires-Dist: requests>=2.28.0
+Dynamic: license-file

{drift_ml-0.1.10 → drift_ml-0.2.1}/drift_ml.egg-info/SOURCES.txt

@@ -1,3 +1,4 @@
+LICENSE
 README.md
 pyproject.toml
 drift/__init__.py

{drift_ml-0.1.10 → drift_ml-0.2.1}/pyproject.toml

@@ -4,13 +4,17 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "drift-ml"
-version = "0.1.10"
-description = "Terminal-first AutoML CLI - chat-based ML engineer"
+version = "0.2.1"
+description = "drift — terminal-first, chat-based AutoML. Open source. No tokens. No auth."
 requires-python = ">=3.10"
 dependencies = [
     "requests>=2.28.0",
 ]
 
+[project.urls]
+Homepage = "https://github.com/lakshitsachdeva/intent2model"
+Repository = "https://github.com/lakshitsachdeva/intent2model"
+
 [project.scripts]
 drift = "drift.__main__:main"
 

drift_ml-0.1.10/PKG-INFO

@@ -1,6 +0,0 @@
-Metadata-Version: 2.4
-Name: drift-ml
-Version: 0.1.10
-Summary: Terminal-first AutoML CLI - chat-based ML engineer
-Requires-Python: >=3.10
-Requires-Dist: requests>=2.28.0

drift_ml-0.1.10/README.md

@@ -1,186 +0,0 @@
-# Intent2Model 🚀
-
-LLM-Guided AutoML Agent - Upload a CSV, chat with the AI, get a trained model.
-
-## ⚡ Quick Start (Easiest Way)
-
-```bash
-# Make scripts executable (first time only)
-chmod +x start.sh stop.sh
-
-# Start everything
-./start.sh
-```
-
-Then open **http://localhost:3000** in your browser!
-
-To stop: Press `Ctrl+C` or run `./stop.sh`
-
----
-
-## 📖 How to Use
-
-### 1. Upload CSV
-- Drag & drop your CSV file or click "choose file"
-- System analyzes it automatically
-
-### 2. Train Model
-Just type a column name in the chat:
-- **"variety"** → trains model to predict "variety"
-- **"price"** → trains model to predict "price"
-- Or any column name from your dataset
-
-### 3. View Results
-- **"report"** → shows beautiful charts and metrics
-- **"show me results"** → displays model performance
-
-### 4. Make Predictions
-- **"predict"** or **"can you predict for me?"** → starts prediction flow
-- Provide feature values: **"sepal.length: 5.1, sepal.width: 3.5"**
-
----
-
-## 💬 Example Conversation
-
-```
-You: [uploads iris.csv]
-AI: ✓ analyzed your dataset • 150 rows • 5 columns
-AI: suggested targets: variety, sepal.length, sepal.width
-
-You: variety
-AI: 🚀 training model to predict "variety"...
-AI: ✅ model trained successfully!
-AI: accuracy: 1.000 • best model: RandomForest
-AI: [shows charts]
-
-You: report
-AI: [shows detailed charts: metrics, feature importance, CV scores]
-
-You: predict
-AI: sure! i need: sepal.length, sepal.width, petal.length, petal.width
-
-You: sepal.length: 5.1, sepal.width: 3.5, petal.length: 1.4, petal.width: 0.2
-AI: 🎯 prediction: Setosa
-AI: probabilities: Setosa 99.8%, Versicolor 0.2%, Virginica 0.0%
-```
-
----
-
-## 🛠️ Manual Setup (Alternative)
-
-### Backend
-**Important:** Always run uvicorn from inside the `backend/` folder. Running from project root will fail with "Could not import module main".
-
-```bash
-cd backend
-pip install -r ../requirements.txt
-# Optional: set API key in .env or export GEMINI_API_KEY=your_key
-python3 -m uvicorn main:app --host 0.0.0.0 --port 8000 --reload
-```
-
-Or use the helper script (from project root):
-```bash
-chmod +x backend/run.sh
-./backend/run.sh
-```
-
-**Backend nahi chal raha?**
-- **"Could not import module main"** → You're in the wrong folder. Run `cd backend` first, then `python3 -m uvicorn main:app --host 0.0.0.0 --port 8000`.
-- **"Address already in use" / port 8000** → Free the port: `lsof -ti:8000 | xargs kill -9`, then start again.
-- **Dependencies missing** → From project root: `pip install -r requirements.txt` (or use `./start.sh` — it creates a venv and installs deps).
-
-### Frontend (New Terminal)
-```bash
-cd frontend
-npm install
-npm run dev
-```
-
-Visit **http://localhost:3000**
-
----
-
-## 🎨 Features
-
-- 📊 **Beautiful Charts**: Metrics, feature importance, CV scores
-- 🎨 **Extravagant UI**: Gradient colors, smooth animations
-- 🤖 **LLM-Powered**: Gemini AI generates optimal pipelines
-- 🔮 **Smart Predictions**: Chat-based prediction interface
-- 📈 **Model Comparison**: Tries multiple models, picks best
-- ⚡ **Auto-Detection**: Automatically detects task type and metrics
-
----
-
-## 📝 Requirements
-
-- Python 3.10+
-- Node.js 18+
-- npm/yarn
-
-Install dependencies:
-```bash
-# Backend
-pip install -r requirements.txt
-
-# Frontend
-cd frontend
-npm install
-```
-
----
-
-## 🐛 Troubleshooting
-
-**Services not starting?**
-- Check ports: `lsof -i :8000` and `lsof -i :3000`
-- Check logs: `tail -f backend.log` or `tail -f frontend.log`
-
-**Training errors?**
-- Make sure CSV has valid data
-- Check that target column exists
-- Try a different column name
-
----
-
-## drift — Terminal-first CLI
-
-**drift** by Lakshit Sachdeva. Terminal-first, chat-based AutoML — same engine as the web UI. No commands to memorize.
-
-### Exactly what to do (any computer)
-
-1. **Install drift** (pick one):
-   ```bash
-   npm install -g drift-ml
-   ```
-   or:
-   ```bash
-   pipx install drift
-   ```
-
-2. **Run drift:**
-   ```bash
-   drift
-   ```
-   You’ll see the welcome and step-by-step instructions in the terminal.
-
-3. **Engine** — On first run the CLI downloads and starts the drift engine locally (or set `DRIFT_BACKEND_URL` to a running engine). You need an LLM: Gemini CLI, Ollama, or another local LLM.
-
-4. **In drift:** type `load path/to/your.csv`, then chat (e.g. `predict price`, `try something stronger`). Type `quit` to exit.
-
-drift shows you the rest when you run it.
-
-### Install (details)
-
-- **Local-first** — Same engine as the web app; planning and training run on your machine.
-- **Chat-based**: e.g. `load iris.csv`, `predict price`, `try something stronger`, `why is accuracy capped`.
-- **Engine** runs locally (CLI auto-starts it or use `DRIFT_BACKEND_URL`). Web UI can be hosted on Vercel.
-
----
-
-## 📚 More Info
-
-See `HOW_TO_USE.md` for detailed instructions and examples.
-
----
-
-**That's it! Just run `./start.sh` and start chatting! 🎉**

drift_ml-0.1.10/drift_ml.egg-info/PKG-INFO

@@ -1,6 +0,0 @@
-Metadata-Version: 2.4
-Name: drift-ml
-Version: 0.1.10
-Summary: Terminal-first AutoML CLI - chat-based ML engineer
-Requires-Python: >=3.10
-Requires-Dist: requests>=2.28.0