ZYJJ Toolset

website·Indexed 31 days ago

https://zyjj.cc

A comprehensive collection of AI-powered productivity tools and media utilities. Features include GPT-based content generators for articles, poems, and social media (Xiaohongshu), text analysis tools for summarization and scoring, and media processing utilities such as Bilibili video/subtitle parsing, NCM to MP3 conversion, voice synthesis, and image enhancement.

Tokens
26.4K
Snippets
154
Records
251
Agent score
51%

What's inside ZYJJ

  1. Overview of Zhiyou Editing (智游剪辑) capabilities

    Zhiyou Editing (智游剪辑) is an AI-powered video editing tool designed to simplify video creation. It provides a wide range of utilities across video processing, subtitle management, AI content generation, and image manipulation.
  2. Use subtitle generation and voice cloning in version 1.2.0

    Version 1.2.0 introduces two new AI-powered capabilities: subtitle generation powered by Whisper and voice cloning powered by OpenVoice.
  3. Use B站小工具 for Bilibili content extraction

    B站小工具 is a Python SDK-based utility that provides a web interface for several Bilibili-related tasks, including cover image parsing, subtitle downloading, danmaku (bullet chat) downloading, video summarization, and comment word cloud generation.
    # Project source code available at:
    https://github.com/zyjj-cc/open-demo/tree/main?tab=readme-ov-file#b%E7%AB%99%E5%B0%8F%E5%B7%A5%E5%85%B7
  4. Download Zhiyou Editing (智游剪辑) for Desktop and Mobile

    Zhiyou Editing provides installers for multiple platforms. Ensure you select the version compatible with your hardware architecture.
  5. Use video editing and data visualization features in version 1.8.0

    Starting from version 1.8.0, the software officially supports video editing. New capabilities include a dedicated 'Creation' (创作) section featuring a full video editor, data visualization video creation tools, and the ability to separate accompaniment from vocals.
  6. Enhance image clarity and quality

    Use the Image Enhancement tool to improve the clarity of low-resolution photos. This operation costs 1 credit per request and supports both synchronous and asynchronous calls.
    from zyjj_open_sdk import Client, FileObject
    
    client = Client('sk-xxxx')
    res = client.image.image_enhance(img=FileObject.from_path("people.png")).execute()
    print(res.img_url)
  7. Generate a word cloud from Bilibili video comments

    The Bilibili Comment Word Cloud tool allows users to visualize the frequency and importance of keywords within a video's comment section to quickly identify main themes. To generate a word cloud, users can use the web version or the 'Zhiyou Jianji' (智游剪辑) desktop application. In the application, search for 'B站评论词云', paste the Bilibili video URL, and click 'Generate' (立即生成).

    Web Version: https://app.zyjj.cc/tool/66b7316e7b8d6d449038fb7c Desktop Protocol: zyjj://tool/66b7316e7b8d6d449038fb7c

  8. Upload large files using multipart upload

    For large files, use the multipart upload process to ensure reliability. The process involves three steps: initializing the upload to get an upload_id, uploading the file in chunks (parts) using provided signed URLs, and completing the upload by submitting the list of part ETags.
    import requests
    import os
    
    base = "https://api.zyjj.cc/api/v1"
    sk = "sk-xxx" # Replace with your secret key
    path = "test.mp4"
    
    # 1. Initialize multipart upload
    file_size = os.path.getsize(path)
    init_res = requests.post(
        f"{base}/open/file/multipart/init",
        json={"file_name": os.path.basename(path), "file_size": file_size},
        headers={"Authorization": sk}
    ).json()
    
    if init_res["code"] != 0:
        raise Exception(init_res["msg"])
    
    upload_id = init_res["data"]
    
    # 2. Upload chunks
    with open(path, "rb") as f:
        part_list = []
        chunk_size = 1024 * 1024 # 1MB chunks
        part_num = 1
        while True:
            part_data = f.read(chunk_size)
            if not part_data:
                break
            
            # Get authorization for this specific part
            auth_res = requests.post(
                f"{base}/open/file/multipart/part",
                json={"upload_id": upload_id, "part_num": part_num},
                headers={"Authorization": sk}
            ).json()
            
            if auth_res["code"] != 0:
                raise Exception(auth_res["msg"])
            
            upload_info = auth_res["data"]
            # Upload actual data to the provided URL
            res = requests.put(
                upload_info["url"],
                data=part_data,
                headers=upload_info["header"]
            )
            res.raise_for_status()
            
            # Store part number and etag for completion
            part_list.append({"part_num": part_num, "etag": res.headers.get("etag")})
            part_num += 1
    
    # 3. Complete the upload
    complete_res = requests.post(
        f"{base}/open/file/multipart/complete",
        json={"upload_id": upload_id, "part_list": part_list},
        headers={"Authorization": sk}
    ).json()
    
    if complete_res["code"] != 0:
        raise Exception(complete_res["msg"])
    
    print(complete_res["data"])
  9. Generate voiceovers using Microsoft Text-to-Speech in Zhiyou Editing

    The Microsoft Text-to-Speech tool in Zhiyou Editing allows users to convert text into spoken audio for video dubbing. The average speaking rate is approximately 310 characters per minute. To use the tool, open the Zhiyou Editing application, search for '语音合成' (Voice Synthesis), select a speaker, enter the text, and click '立即合成' (Synthesize Now).
  10. Generate multilingual subtitles from audio

    Use the multilingual subtitle recognition service to generate subtitles for various languages. This is an asynchronous process that costs 10 points per minute. The resulting subtitle link is valid for 6 hours.

    • Entity ID: 935badd005f9c08c29d01b85
    • Input Parameters:
      • audio (Required): Audio file in .mp3 or .wav format.
      • language (Required): Language code. Options include auto (automatic, lower accuracy), zh (Chinese), en (English), ja (Japanese), ko (Korean), fr (French), de (German), and many others (e.g., ar, ru, es, vi).
    • Output: A URL to the generated subtitle file (subtitle).
    from zyjj_open_sdk import Client, FileObject
    
    client = Client('sk-xxxx')
    res = client.subtitle.subtitle_generate_multilingual(audio=FileObject.from_path("test.mp3"), language="zh").execute_async().wait(lambda i: print(f"progres: {i}"))
    print(res.subtitle)
  11. Extract album art from MP3 files using Zhiyou Editing

    To extract the embedded cover image from an MP3 music file, use the 'Music Cover Extraction' (音乐封面提取) tool within the Zhiyou Editing (智游剪辑) application. The process involves searching for the tool by name and uploading the target MP3 file.