cutted 0.3.1__py3-none-any.whl → 0.3.3__py3-none-any.whl

cutted/__init__.py

@@ -1,2 +1,2 @@
-__version__ = "0.3.1"
+__version__ = "0.3.3"
 __author__ = "simon0302010"

cutted/app.py

@@ -25,6 +25,8 @@ class CuttedApp:
         self.canvas = None
         self.cursor_line = None
         self.last_slider_update = 0
+        self.slider_value = 0
+        self.playback_start_time = 0
         self.is_playing = False
         self.last_states = []
         self.setup_ui()
@@ -58,11 +60,19 @@ class CuttedApp:
         if whisper_support:
             self.use_transcript_checkbox = customtkinter.CTkCheckBox(
                 self.root,
-                text="Give Gemini a transcript (very slow)",
+                text="Send transcript to Gemini (slower, more accurate)",
                 text_color="#888888",
                 font=("Arial", 12)
             )
             self.use_transcript_checkbox.place(relx=0.0, rely=1.0, anchor="w", y=-12)
+            
+        self.use_audio_checkbox = customtkinter.CTkCheckBox(
+            self.root,
+            text="Send audio to Gemini (buggy)",
+            text_color="#888888",
+            font=("Arial", 12)
+        )
+        self.use_audio_checkbox.place(relx=1.0, rely=1.0, anchor="e", y=-12)
 
         self.play_button = customtkinter.CTkButton(self.root, text="Play", command=self.play_audio, width=50)
         self.play_button.place(relx=0.3, rely=1.0, anchor="s", y=-30)
@@ -105,7 +115,7 @@ class CuttedApp:
         self.canvas = FigureCanvasTkAgg(fig, master=self.plot_frame)
         self.canvas.draw()
         
-        self.audio_lenght = int(round(self.AudioProcessor.get_lenght()))
+        self.audio_lenght = int(round(self.AudioProcessor.get_length()))
         
         slider_width = self.root.winfo_width() - 40
         self.slider = customtkinter.CTkSlider(
@@ -145,11 +155,16 @@ class CuttedApp:
             return
 
         start_time = self.slider.get() if hasattr(self, 'slider') else 0
+        self.playback_start_time = start_time
         self.AudioProcessor.play_audio(start_time)
 
     def stop_audio(self):
-        self.AudioProcessor.stop_audio()
+        rel_pos = self.AudioProcessor.stop_audio()
         self.is_playing = False
+        abs_pos = self.playback_start_time + rel_pos
+        self.slider.set(abs_pos)
+        self.set_cursor(abs_pos)
+        print_info(f"Absolute position in audio: {abs_pos:.2f}s")
         
     def export_audio(self):
         if not hasattr(self.AudioProcessor, "audio") or self.AudioProcessor.audio is None:
@@ -176,6 +191,7 @@ class CuttedApp:
             print_success(f"Audio exported to {save_path}")
             
     def send_prompt(self):
+        print(self.AudioProcessor.get_waveform_summary())
         self.save_state()
 
         if not hasattr(self.AudioProcessor, "audio") or self.AudioProcessor.audio is None:
@@ -183,32 +199,41 @@ class CuttedApp:
             return
         
         text = self.entry.get()
-        full_prompt = f"You are a audio editing AI. You are controllable via natural language and editing a audio file. The audio file is {round(self.AudioProcessor.get_lenght())}s long."
-        if whisper_support:
-            if self.use_transcript_checkbox.get():
-                if not self.whisper:
-                    messagebox.showinfo("Info", "Loading Whisper model. This may take a few minutes depending on your internet connection. See the progress in your command line. If this window appears to be frozen, the transcription is running.")
-                    self.whisper = transcribe.Whisper()
-                transcript = self.whisper.transcribe(self.AudioProcessor.audio_path)
-                full_prompt += f"\nThis is a transcript with per word timestamps of the audio:\n{transcript}"
-        full_prompt += f"\n\nUser Prompt: {text}"
-        self.entry.delete(0, "end")
-        
-        function_call, text_result = self.gemini.generate(full_prompt)
-        
-        if function_call:
-            print_info(f"Gemini called {function_call.name}")
-            if function_call.name == "cut_audio":
-                print_info("Cut function called")
-                args = function_call.args
-                result = self.AudioProcessor.cut(args["start"], args["end"])
-                if not result:
-                    messagebox.showerror("Error", "Please try again.")
-            self.update_plot()
-        elif text_result:
-            messagebox.showerror("Error", text_result.strip())
-        else:
-            print_fail("Gemini returned no data")
+        if text.strip():
+            full_prompt = f"You are a audio editing AI. You are controllable via natural language and editing a audio file. The audio file is {round(self.AudioProcessor.get_length())}s long. The cursor of the user is currently at {self.slider_value}s."
+            full_prompt += "\nHere is a the waveform samples of the audio. You can use them to determine silent parts, loud parts, silences, beats and much more.\nYou are forced to used these if the user requires you to cut out silent of quiet parts for example."
+            full_prompt += "\nAll of your tools should be enough to fullfill almost every task.\nNEVER ASK FOR CONFIRMATION FROM THE USER. DO EVERYTHING!"
+            full_prompt += f"\n{self.AudioProcessor.get_waveform_summary()}\n"
+            if whisper_support:
+                if self.use_transcript_checkbox.get():
+                    if not self.whisper:
+                        messagebox.showinfo("Info", "Loading Whisper model. This may take a few minutes depending on your internet connection. See the progress in your command line. If this window appears to be frozen, the transcription is running. Press OK to continue.")
+                        self.whisper = transcribe.Whisper()
+                    transcript = self.whisper.transcribe(self.AudioProcessor.audio_path)
+                    full_prompt += f"\nThis is a transcript with per word timestamps of the audio:\n{transcript}"
+                    full_prompt += "\nThe transcript likely has issues. If you need infos about some words they might just be misspelled in the audio."
+            full_prompt += f"\n\nUser Prompt: {text}"
+            self.entry.delete(0, "end")
+            
+            if self.use_audio_checkbox.get():
+                function_call, text_result = self.gemini.generate(full_prompt, audio_base64=self.AudioProcessor.get_audio_base64())
+            else:
+                function_call, text_result = self.gemini.generate(full_prompt)
+            
+            if function_call:
+                print_info(f"Gemini called {function_call.name}")
+                if function_call.name == "cut_audio":
+                    print_info("Cut function called")
+                    args = function_call.args
+                    result = self.AudioProcessor.cut(args["start"], args["end"])
+                    if not result:
+                        messagebox.showerror("Error", "Please try again.")
+                self.update_plot()
+            elif text_result:
+                messagebox.showerror("Error", text_result.strip())
+            else:
+                messagebox.showerror("Error", "Gemini returned no data")
+                print_fail("Gemini returned no data")
 
     def save_state(self):
         if hasattr(self.AudioProcessor, "audio") and self.AudioProcessor.audio is not None:

cutted/core/audio_processor.py

@@ -4,9 +4,8 @@ from .logger import *
 import numpy as np
 from matplotlib.figure import Figure
 import pygame
+import base64
 import io
-import threading
-import time
 
 class AudioProcessor:
     def __init__(self):
@@ -63,7 +62,21 @@ class AudioProcessor:
         
         return fig
     
-    def get_lenght(self):
+    def get_waveform_summary(self):
+        num_samples = round(self.get_length())
+        if self.audio is None:
+            return "No audio loaded."
+        samples = np.array(self.audio.get_array_of_samples())
+        if self.audio.channels == 2:
+            samples = samples.reshape((-1, 2))
+            samples = samples.mean(axis=1)
+        samples = samples / np.max(np.abs(samples))
+        indices = np.linspace(0, len(samples)-1, num_samples).astype(int)
+        summary = samples[indices]
+        return f"Waveform samples (normalized, {num_samples} points):\n" + \
+            " ".join(f"{x:.2f}" for x in summary)
+    
+    def get_length(self):
         self.duration = self.audio.duration_seconds
         self.duration = round(self.duration, 2)
         return self.duration
@@ -71,17 +84,25 @@ class AudioProcessor:
     def cut(self, start, end):
         if len(start) == len(end):
             if len(start) == 1:
-                print_info(f"Cutting from {start[0]} to {end[0]}")
-                start_ms = round(start[0] * 1000)
-                end_ms = round(end[0] * 1000)
+                single_start = max(0, start[0])
+                single_end = max(0, end[0])
+                if single_end <= single_start:
+                    print_fail("End time must be greater than start time.")
+                    return False
+                print_info(f"Cutting from {single_start} to {single_end}")
+                start_ms = round(single_start * 1000)
+                end_ms = round(single_end * 1000)
                 self.audio = self.audio[:start_ms] + self.audio[end_ms:]
                 return True
             else:
                 time_sets = list(zip(start, end))
                 subtract_time = 0
                 for single_start, single_end in time_sets:
-                    single_start = single_start - subtract_time
-                    single_end = single_end - subtract_time
+                    single_start = max(0, single_start - subtract_time)
+                    single_end = max(0, single_end - subtract_time)
+                    if single_end <= single_start:
+                        print_fail("End time must be greater than start time.")
+                        continue
                     print_info(f"Cutting from {single_start} to {single_end}")
                     start_ms = round(single_start * 1000)
                     end_ms = round(single_end * 1000)
@@ -124,11 +145,15 @@ class AudioProcessor:
     def stop_audio(self):
         try:
             if pygame.mixer.get_init():
+                pos_ms = pygame.mixer.music.get_pos()
+                pos_sec = pos_ms / 1000 if pos_ms >= 0 else 0
                 pygame.mixer.music.stop()
                 self.is_playing_var = False
-                print_info("Audio playback stopped")
+                print_info(f"Audio playback stopped at {pos_sec:.2f}s")
+                return pos_sec
         except Exception as e:
             print_warn(f"Error stopping audio: {e}")
+            return 0
             
     def is_playing(self):
         try:
@@ -143,11 +168,19 @@ class AudioProcessor:
             return None
         
         return {
-            "duration": self.get_lenght(),
+            "duration": self.get_length(),
             "channels": self.audio.channels,
             "frame_rate": self.audio.frame_rate,
             "sample_width": self.audio.sample_width
         }
         
     def export_audio(self, path, format: str = "mp3"):
-        self.audio.export(path, format=format)
+        self.audio.export(path, format=format)
+        
+    def get_audio_base64(self):
+        buffer = io.BytesIO()
+        self.audio.export(buffer, format="mp3")
+        buffer.seek(0)
+        audio_bytes = buffer.read()
+        audio_base64 = base64.b64encode(audio_bytes)
+        return audio_base64

cutted/core/gemini.py

@@ -1,5 +1,6 @@
 import os
 import sys
+import base64
 from google import genai
 from google.genai import types
 
@@ -15,13 +16,21 @@ class GeminiClient:
             api_key=GEMINI_API_KEY,
         )
     
-    def generate(self, prompt: str, model: str = "gemini-2.0-flash"):
+    def generate(self, prompt: str, model: str = "gemini-2.0-flash", audio_base64 = None):
+        parts=[
+            types.Part.from_text(text=prompt),
+        ]
+        
+        if audio_base64:
+            parts.append(types.Part.from_bytes(
+                mime_type="audio/mpeg",
+                data=base64.b64decode(audio_base64)
+            ))
+        
         contents = [
             types.Content(
                 role="user",
-                parts=[
-                    types.Part.from_text(text=prompt),
-                ],
+                parts=parts
             ),
         ]
         tools = [

cutted-0.3.3.dist-info/METADATA

@@ -0,0 +1,55 @@
+Metadata-Version: 2.4
+Name: cutted
+Version: 0.3.3
+Summary: AI-powered audio editor controllable via natural language.
+Author-email: simon0302010 <simon0302010@gmail.com>
+License-Expression: GPL-3.0
+Project-URL: Homepage, https://github.com/simon0302010/Cutted
+Requires-Python: <=3.13,>=3.9
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: customtkinter
+Requires-Dist: matplotlib
+Requires-Dist: numpy
+Requires-Dist: pydub
+Requires-Dist: pygame
+Requires-Dist: google-genai
+Requires-Dist: python-dotenv
+Provides-Extra: whisper
+Requires-Dist: whisper-timestamped; extra == "whisper"
+Dynamic: license-file
+
+![PyPI](https://img.shields.io/pypi/v/cutted?color=blue)
+![PyPI - License](https://img.shields.io/pypi/l/lyriks-video)
+![Hackatime](https://hackatime-badge.hackclub.com/U08HC7N4JJW/Cutted)
+
+# Cutted
+
+AI-powered audio editor controlled by natural language 🚀
+
+Let AI handle your audio editing with simple commands:
+- Automatically detect quiet or loud parts
+- Transcribe audio (if Whisper is installed)
+- Cut, trim, or adjust volume for specific segments
+
+## Installation
+
+1. **Install system dependencies (e.g., FFmpeg).**  
+2. **Install Cutted**  
+```bash
+pip install cutted
+```
+**With Whisper support**  
+```bash
+pip install cutted[whisper]
+```
+  
+## Usage
+
+1. **Launch the app:**  
+   ```bash
+   python -m cutted
+   ```
+2. **Load an audio file** (MP3, WAV, etc.)  
+3. **Play, Cut, Undo** – all from the GUI or with text commands ✂️
+4. **Export** as MP3 or WAV  

cutted-0.3.3.dist-info/RECORD

@@ -0,0 +1,12 @@
+cutted/__init__.py,sha256=F1mzO6qI2gD6d_DQsjZItLuIfNfa1Te5KwCMThWDQT4,49
+cutted/__main__.py,sha256=lYGLgtIZ_vGZIJmWG6ZQoqOdyOJnaWEA4NBn5Rc7Q8E,61
+cutted/app.py,sha256=HaJ3yu8-WVDGKZmFcTVG4CK8q13eHyUQaIyOWMZWV68,10764
+cutted/core/audio_processor.py,sha256=gKBJ1wpdrX2IozmXDzW7MoFVp9uq0Pb3ezz9R3Ahmnw,6691
+cutted/core/gemini.py,sha256=yHsQXk4tDHcW5qJBuL2LpPvdChimxlmEbu76BvsoeY4,3108
+cutted/core/logger.py,sha256=AjqrgW2LV9HdPkPQ8oOmyd9lWzVSIg46r74ILR7mVHo,585
+cutted/core/transcribe.py,sha256=0e7aCva4y6D-gKe1xw5HT9VoFgbvHGgV6utn12r8wXA,986
+cutted-0.3.3.dist-info/licenses/LICENSE,sha256=OXLcl0T2SZ8Pmy2_dmlvKuetivmyPd5m1q-Gyd-zaYY,35149
+cutted-0.3.3.dist-info/METADATA,sha256=ehEIEm5qz3wVqBY74Itpwooe4SYOFpXISXUMZO7LVh4,1503
+cutted-0.3.3.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+cutted-0.3.3.dist-info/top_level.txt,sha256=PL6glZvzRyKWCDn5aoYI9uH8HlEA5Qd_XFJowJKARYI,7
+cutted-0.3.3.dist-info/RECORD,,

cutted-0.3.1.dist-info/METADATA

@@ -1,68 +0,0 @@
-Metadata-Version: 2.4
-Name: cutted
-Version: 0.3.1
-Summary: AI-powered audio editor controllable via natural language.
-Author-email: simon0302010 <simon0302010@gmail.com>
-License-Expression: GPL-3.0
-Project-URL: Homepage, https://github.com/simon0302010/Cutted
-Requires-Python: <=3.13,>=3.9
-Description-Content-Type: text/markdown
-License-File: LICENSE
-Requires-Dist: customtkinter
-Requires-Dist: matplotlib
-Requires-Dist: numpy
-Requires-Dist: pydub
-Requires-Dist: pygame
-Requires-Dist: google-genai
-Requires-Dist: python-dotenv
-Provides-Extra: whisper
-Requires-Dist: whisper-timestamped; extra == "whisper"
-Dynamic: license-file
-
-# Cutted
-AI-powered audio editor controllable via natural language
-
-Note: This app is currently not functional.
-
-# Installation
-
-Install dependencies:
-
-**Debian:**
-```bash
-sudo apt update
-sudo apt install ffmpeg git
-```
-
-**Arch Linux:**
-```bash
-sudo pacman -Syu ffmpeg git
-```
-
-Clone the repository:
-
-```bash
-git clone https://github.com/simon0302010/Cutted.git
-cd Cutted
-```
-
-Create a virtual environment:
-
-```bash
-python -m venv venv
-source venv/bin/activate
-```
-
-Install the package:
-
-```bash
-pip install .
-```
-
-# Usage
-
-Run the package:
-
-```bash
-python -m cutted
-```

cutted-0.3.1.dist-info/RECORD

@@ -1,12 +0,0 @@
-cutted/__init__.py,sha256=Q7ZgbkmUOc-Gzy42zzZcKnzKxdZ023pITyXAC0BDBlM,49
-cutted/__main__.py,sha256=lYGLgtIZ_vGZIJmWG6ZQoqOdyOJnaWEA4NBn5Rc7Q8E,61
-cutted/app.py,sha256=xf5LFHdPLz1WNP05eRruHX2hKnaei8CUrqzVSODweAM,8952
-cutted/core/audio_processor.py,sha256=7-XCuPPTlozeuaD2LqyzwRGinu0NvowTLbAh2X4XJ98,5182
-cutted/core/gemini.py,sha256=Ts_EbC1-rO9jIsdSlzKcmjLVS1o663GmfTdzmix12kE,2872
-cutted/core/logger.py,sha256=AjqrgW2LV9HdPkPQ8oOmyd9lWzVSIg46r74ILR7mVHo,585
-cutted/core/transcribe.py,sha256=0e7aCva4y6D-gKe1xw5HT9VoFgbvHGgV6utn12r8wXA,986
-cutted-0.3.1.dist-info/licenses/LICENSE,sha256=OXLcl0T2SZ8Pmy2_dmlvKuetivmyPd5m1q-Gyd-zaYY,35149
-cutted-0.3.1.dist-info/METADATA,sha256=6plv-IaqUXQeeAnChYaMBjYSLh8FBTAtvyQqZkRTuuA,1201
-cutted-0.3.1.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-cutted-0.3.1.dist-info/top_level.txt,sha256=PL6glZvzRyKWCDn5aoYI9uH8HlEA5Qd_XFJowJKARYI,7
-cutted-0.3.1.dist-info/RECORD,,