ai-screenshooter 1.5.0__tar.gz → 1.7.1__tar.gz

{ai_screenshooter-1.5.0 → ai_screenshooter-1.7.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ai-screenshooter
-Version: 1.5.0
+Version: 1.7.1
 Summary: A CLI tool to capture and send AI-powered screenshots
 Home-page: https://github.com/tech4vision/ai-screenshoter
 Author: Last Shot AI
@@ -14,6 +14,10 @@ Requires-Dist: requests
 Requires-Dist: Pillow
 Requires-Dist: pygetwindow
 Requires-Dist: pyperclip
+Requires-Dist: sounddevice
+Requires-Dist: soundfile
+Requires-Dist: numpy
+Requires-Dist: faster-whisper
 Dynamic: author
 Dynamic: author-email
 Dynamic: classifier

{ai_screenshooter-1.5.0 → ai_screenshooter-1.7.1}/ai_screenshooter.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ai-screenshooter
-Version: 1.5.0
+Version: 1.7.1
 Summary: A CLI tool to capture and send AI-powered screenshots
 Home-page: https://github.com/tech4vision/ai-screenshoter
 Author: Last Shot AI
@@ -14,6 +14,10 @@ Requires-Dist: requests
 Requires-Dist: Pillow
 Requires-Dist: pygetwindow
 Requires-Dist: pyperclip
+Requires-Dist: sounddevice
+Requires-Dist: soundfile
+Requires-Dist: numpy
+Requires-Dist: faster-whisper
 Dynamic: author
 Dynamic: author-email
 Dynamic: classifier

{ai_screenshooter-1.5.0 → ai_screenshooter-1.7.1}/ai_screenshooter.egg-info/requires.txt

@@ -3,3 +3,7 @@ requests
 Pillow
 pygetwindow
 pyperclip
+sounddevice
+soundfile
+numpy
+faster-whisper

{ai_screenshooter-1.5.0 → ai_screenshooter-1.7.1}/ai_screenshot.py

@@ -1,4 +1,5 @@
 import argparse
+import json
 import os
 import sys
 import signal
@@ -6,6 +7,7 @@ import logging
 import atexit
 import time
 import subprocess
+import threading
 import requests
 import pygetwindow as gw
 import pyperclip
@@ -17,9 +19,17 @@ from pynput import keyboard
 # Constants
 PID_FILE = Path.home() / ".ai-screenshooter.pid"
 LOG_FILE = Path.home() / ".ai-screenshooter.log"
+META_FILE = Path.home() / ".ai-screenshooter.meta.json"
 SCREENSHOT_DIR = Path.home() / ".ai-screenshooter" / "screenshots"
+AUDIO_DIR = Path.home() / ".ai-screenshooter" / "audio"
 TIMEOUT_SECONDS = 5 * 60 * 60  # 5 hours
 
+# Audio recording constants
+SAMPLE_RATE = 16000  # Whisper expects 16kHz
+CHANNELS = 1  # Mono audio
+WHISPER_MODEL = "base"  # Options: tiny, base, small, medium, large
+DOUBLE_TAP_THRESHOLD = 0.5  # 500ms window for double-tap
+
 # Server URLs
 PROD_URL = "https://service.tech4vision.net/ai-management-service/api/v1/sessions/code-challenge"
 LOCAL_URL = "http://localhost:8082/api/v1/sessions/code-challenge"
@@ -31,6 +41,13 @@ API_URL = None
 current_keys = set()
 logger = logging.getLogger("ai-screenshooter")
 
+# Voice recording state
+is_recording = False
+audio_thread = None
+audio_data = []
+whisper_model = None  # Lazy-loaded on first use
+last_esc_time = 0  # For double-tap detection
+
 if sys.platform == "win32":
     import ctypes
     from ctypes import Structure, c_long
@@ -82,6 +99,31 @@ def cleanup_pid_file():
             PID_FILE.unlink()
     except Exception:
         pass
+    try:
+        if META_FILE.exists():
+            META_FILE.unlink()
+    except Exception:
+        pass
+
+
+def write_meta_file(server_mode, server_url):
+    """Write process metadata for status command."""
+    meta = {
+        "started_at": time.time(),
+        "server_mode": server_mode,
+        "server_url": server_url,
+    }
+    META_FILE.write_text(json.dumps(meta))
+
+
+def read_meta_file():
+    """Read process metadata, return None if invalid."""
+    if not META_FILE.exists():
+        return None
+    try:
+        return json.loads(META_FILE.read_text())
+    except (ValueError, IOError):
+        return None
 
 
 # ============ Process Management ============
@@ -299,11 +341,196 @@ def send_clipboard_text():
         logger.error(f"Error sending clipboard text: {e}")
 
 
+# ============ Voice Recording Functions ============
+
+def get_whisper_model():
+    """Lazy-load Whisper model on first use."""
+    global whisper_model
+    if whisper_model is None:
+        try:
+            from faster_whisper import WhisperModel
+            logger.info(f"Loading Whisper model '{WHISPER_MODEL}' (first time may download ~74MB)...")
+            whisper_model = WhisperModel(WHISPER_MODEL, device="cpu", compute_type="int8")
+            logger.info("Whisper model loaded successfully.")
+        except Exception as e:
+            logger.error(f"Failed to load Whisper model: {e}")
+            return None
+    return whisper_model
+
+
+def record_audio():
+    """Record audio from microphone in a separate thread."""
+    global audio_data, is_recording
+    import sounddevice as sd
+
+    audio_data = []
+
+    def audio_callback(indata, frames, time_info, status):
+        if status:
+            logger.warning(f"Audio status: {status}")
+        if is_recording:
+            audio_data.append(indata.copy())
+
+    try:
+        with sd.InputStream(samplerate=SAMPLE_RATE, channels=CHANNELS,
+                           callback=audio_callback, dtype='float32'):
+            while is_recording:
+                sd.sleep(100)  # Sleep 100ms, check if still recording
+    except Exception as e:
+        logger.error(f"Microphone error: {e}")
+
+
+def start_voice_recording():
+    """Start recording audio in a background thread."""
+    global is_recording, audio_thread, audio_data
+
+    if is_recording:
+        return  # Already recording
+
+    logger.info("Voice recording started... (release ESC to stop)")
+    is_recording = True
+    audio_data = []
+
+    audio_thread = threading.Thread(target=record_audio, daemon=True)
+    audio_thread.start()
+
+
+def stop_voice_recording_and_send():
+    """Stop recording, transcribe audio, and send to API."""
+    global is_recording, audio_thread, audio_data
+
+    if not is_recording:
+        return
+
+    logger.info("Voice recording stopped, processing...")
+    is_recording = False
+
+    # Wait for recording thread to finish
+    if audio_thread:
+        audio_thread.join(timeout=1.0)
+
+    # Check if we have audio data
+    if not audio_data:
+        logger.warning("No audio recorded.")
+        return
+
+    # Combine audio chunks
+    try:
+        import numpy as np
+        import soundfile as sf
+
+        audio_array = np.concatenate(audio_data, axis=0)
+
+        # Minimum recording duration check (0.5 seconds)
+        if len(audio_array) < SAMPLE_RATE * 0.5:
+            logger.warning("Recording too short, ignoring.")
+            return
+
+        # Save to temporary file
+        AUDIO_DIR.mkdir(parents=True, exist_ok=True)
+        temp_audio_path = AUDIO_DIR / f"recording_{int(time.time())}.wav"
+
+        sf.write(str(temp_audio_path), audio_array, SAMPLE_RATE)
+        logger.info(f"Audio saved: {temp_audio_path}")
+
+        # Transcribe
+        transcribed_text = transcribe_audio(temp_audio_path)
+
+        if transcribed_text:
+            # Send to API
+            send_transcribed_text(transcribed_text)
+
+    except Exception as e:
+        logger.error(f"Error processing audio: {e}")
+    finally:
+        # Cleanup temp file
+        try:
+            if 'temp_audio_path' in locals() and temp_audio_path.exists():
+                temp_audio_path.unlink()
+        except Exception:
+            pass
+
+
+def transcribe_audio(audio_path):
+    """Transcribe audio file using Whisper."""
+    try:
+        model = get_whisper_model()
+        if model is None:
+            return None
+
+        logger.info("Transcribing audio...")
+        segments, info = model.transcribe(str(audio_path), beam_size=5)
+
+        # Combine all segments
+        text = " ".join([segment.text.strip() for segment in segments])
+
+        if text:
+            logger.info(f"Transcription: {text[:100]}{'...' if len(text) > 100 else ''}")
+        else:
+            logger.warning("Transcription returned empty text.")
+
+        return text
+
+    except Exception as e:
+        logger.error(f"Transcription error: {e}")
+        return None
+
+
+def send_transcribed_text(text):
+    """Send transcribed text to the Code tab API."""
+    if not API_TOKEN:
+        logger.error("No API token provided!")
+        return
+
+    if not text or not text.strip():
+        logger.warning("No text to send.")
+        return
+
+    try:
+        response = requests.post(
+            f"{API_URL}/chat",
+            headers={
+                "Authorization": f"Bearer {API_TOKEN}",
+                "Content-Type": "application/json"
+            },
+            json={"message": text}
+        )
+
+        if response.status_code == 200:
+            logger.info("Transcribed text sent successfully.")
+        else:
+            logger.error(f"Failed to send text: {response.text}")
+    except Exception as e:
+        logger.error(f"Error sending transcribed text: {e}")
+
+
 # ============ Keyboard Handlers ============
 
 def on_press(key):
-    current_keys.add(key)
+    global last_esc_time, is_recording
+
     try:
+        # Double-tap ESC detection for voice recording
+        if key == keyboard.Key.esc:
+            # Ignore repeated key events from holding ESC
+            if keyboard.Key.esc in current_keys:
+                return
+            current_keys.add(key)
+
+            current_time = time.time()
+            time_since_last = current_time - last_esc_time
+
+            if time_since_last < DOUBLE_TAP_THRESHOLD and not is_recording:
+                # Double-tap detected - start recording
+                start_voice_recording()
+
+            last_esc_time = current_time
+
+        # Track non-ESC keys for combo detection
+        else:
+            current_keys.add(key)
+
+        # Other hotkeys (ESC + arrow keys)
         if key == keyboard.Key.down and keyboard.Key.esc in current_keys:
             logger.info("Capturing screenshot...")
             capture_screenshot()
@@ -318,11 +545,18 @@ def on_press(key):
 
 
 def on_release(key):
+    global is_recording
+
     try:
         current_keys.remove(key)
     except KeyError:
         pass
 
+    # Stop voice recording when ESC is released
+    if is_recording and key == keyboard.Key.esc:
+        # Run transcription in background thread to not block keyboard listener
+        threading.Thread(target=stop_voice_recording_and_send, daemon=True).start()
+
 
 # ============ CLI Commands ============
 
@@ -330,19 +564,20 @@ def cmd_start(args):
     """Handle the start command."""
     global API_TOKEN, API_URL
 
-    # If --background flag, spawn a new process and exit
-    if args.background:
-        print("Starting in background mode...")
+    is_daemon = getattr(args, 'daemon', False)
+
+    # Kill any existing instance (unless this is the daemon subprocess itself)
+    if not is_daemon:
         killed = kill_existing_process()
         if killed:
-            print("Killed existing instance.")
+            print("Replaced existing instance.")
 
+    # If --background flag, spawn a new process and exit
+    if args.background:
+        print("Starting in background mode...")
         start_background_process(args.token, args.local)
         return
 
-    # If --daemon flag (internal), this is the actual daemon process
-    is_daemon = getattr(args, 'daemon', False)
-
     if is_daemon:
         # Write PID file
         write_pid_file()
@@ -365,11 +600,16 @@ def cmd_start(args):
     API_URL = LOCAL_URL if args.local else PROD_URL
 
     server_mode = "LOCAL" if args.local else "PRODUCTION"
+
+    # Write metadata for status command
+    write_meta_file(server_mode, API_URL)
+
     logger.info("AI Screenshot CLI started.")
     logger.info(f"Server: {server_mode} ({API_URL})")
     logger.info("Press ESC + Down to capture a screenshot.")
     logger.info("Press ESC + Up to send all stored screenshots.")
     logger.info("Press ESC + Right to send clipboard text to Code tab.")
+    logger.info("Double-tap ESC (hold on 2nd) to record voice and send transcription.")
     if not is_daemon:
         logger.info("Running... (Press Ctrl + C to exit)")
 
@@ -383,11 +623,38 @@ def cmd_status(args):
     pid = get_pid_from_file()
     if pid and is_process_running(pid):
         print(f"ai-screenshooter is running (PID: {pid})")
+
+        meta = read_meta_file()
+        if meta:
+            # Uptime
+            elapsed = time.time() - meta.get("started_at", time.time())
+            hours, remainder = divmod(int(elapsed), 3600)
+            minutes, seconds = divmod(remainder, 60)
+            print(f"  Uptime:  {hours}h {minutes}m {seconds}s")
+
+            # Time remaining
+            remaining = TIMEOUT_SECONDS - elapsed
+            if remaining > 0:
+                rh, rr = divmod(int(remaining), 3600)
+                rm, rs = divmod(rr, 60)
+                print(f"  Expires: {rh}h {rm}m {rs}s remaining")
+
+            # Server
+            print(f"  Server:  {meta.get('server_mode', 'UNKNOWN')} ({meta.get('server_url', '')})")
+
+        print()
+        print("  Listening for hotkeys:")
+        print("    ESC + Down        Capture screenshot")
+        print("    ESC + Up          Send all screenshots")
+        print("    ESC + Right       Send clipboard text to Code tab")
+        print("    Double-tap ESC    Record voice, transcribe and send")
+
         return 0
     else:
         print("ai-screenshooter is not running")
         if PID_FILE.exists():
-            print(f"(stale PID file exists at {PID_FILE})")
+            print(f"(stale PID file found, cleaning up)")
+            cleanup_pid_file()
         return 1
 
 

{ai_screenshooter-1.5.0 → ai_screenshooter-1.7.1}/setup.py

@@ -2,7 +2,7 @@ from setuptools import setup, find_packages
 
 setup(
     name="ai-screenshooter",
-    version="1.5.0",
+    version="1.7.1",
     packages=find_packages(),
     py_modules=["ai_screenshot"],
     install_requires=[
@@ -10,7 +10,11 @@ setup(
         "requests",
         "Pillow",
         "pygetwindow",
-        "pyperclip"
+        "pyperclip",
+        "sounddevice",
+        "soundfile",
+        "numpy",
+        "faster-whisper"
     ],
     entry_points={
         "console_scripts": [