WebUI, PDF Translator for Human web interface

.gitignore

@@ -22,6 +22,7 @@ sdist/
 var/
 .idea/
 .idea
+.cached
 wheels/
 *.egg-info/
 .installed.cfg

README.md

@@ -0,0 +1,81 @@
+# PDF Translator for Human: A PDF Reader/Translator with Local LLM/ChatGPT or Google
+
+## Use Case
+
+There is tons of PDF reader/translator with AI supported. However none of them meets my need. I hope it could run totally on local with local LLMs.
+
+I hope I could read both the original PDF and the translated pages side by side. 
+Also I don't like to translate a 1000 pages long PDF file all at one.
+
+## Features in PDF Translator for Human
+You can read both the original PDF file and the translated content side by side.
+
+The local/remote translation API is invoked on a per-page basis as needed, triggered by page turns during reading.
+
+
+## Supported translators and LLMS:
+* Google Translator (NO need api-key, it it totally free)
+* ChatGPT
+* DeepSeek (Use the OpenAI Compatible endpoint at https://api.deepseek.com/v1)
+* Qwen (Use the OpenAI Compatible endpoint)
+* Local deployed LLMs by ollama, llama.cpp, mlx_lm
+* Other OpenAI Compatible LLMs like GLM/Moonshot etc.
+
+## Start the Web Application for PDF Translator for Human
+
+
+``` bash
+./run_translator_web.sh
+
+# or just start the streamlit application if you have run the previous script:
+streamlit run pdf_translator_web.py
+
+```
+
+## Notes on deploy and start local llm 
+
+### Option 1.Start local llm By mlx_lm (works on Mac Sillicon.)
+
+Here I download aya-expanse-8b 4bit as an example.
+
+``` Bash
+# download mlx models from huggingface to local folder
+git clone https://huggingface.co/mlx-community/aya-expanse-8b-4bit
+
+# install mlx_lm
+pip install mlx_lm
+
+# start the server
+mlx_lm.server --model ./aya-expanse-8b-4bit --port 8080
+
+```
+
+### Option 2. By llama.cpp (Works on CPU/GPU/Mac Machines)
+
+Llama.cpp works on CPU machines and Mac Intel/Sillicon machines, you need 48GB memories for aya-expanse-32b-q4_k_m.gguf.
+
+``` Bash
+# download gguf models from huggingface to local folder
+wget https://hf-mirror.co/bartowski/aya-expanse-32b-GGUF/resolve/main/aya-expanse-32b-Q4_K_M.gguf -O aya-expanse-32b-Q4_K_M.gguf
+
+# download llama.cpp and install llama.cpp
+git clone https://github.com/ggerganov/llama.cpp
+cd llama.cpp
+mkdir -p build && cmake -B build
+cmake --build build --config Release -j 12
+
+# start llama.cpp server
+./llama-server -m ~/models/aya-expanse-32b-Q4_K_M.gguf --port 8080
+
+```
+
+## Snapshot
+
+![PDF Translator for Human](PDF-Translator-for-Human.jpg)
+
+
+## Acknowlegement
+
+https://github.com/nidhaloff/deep-translator
+
+The project is based on the awesome deep-translator.

pdf_translator_web.py

@@ -0,0 +1,291 @@
+import os
+import json
+import hashlib
+from pathlib import Path
+import streamlit as st
+import pymupdf
+from deep_translator import (
+    GoogleTranslator,
+    ChatGptTranslator,
+)
+
+# Constants
+DEFAULT_PAGES_PER_LOAD = 2
+DEFAULT_MODEL = "default_model"
+DEFAULT_API_BASE = "http://localhost:8080/v1"
+
+# Supported translators
+TRANSLATORS = {
+    'google': GoogleTranslator,
+    'chatgpt': ChatGptTranslator,
+}
+
+# Color options
+COLOR_MAP = {
+    "darkred": (0.8, 0, 0),
+    "black": (0, 0, 0),
+    "blue": (0, 0, 0.8),
+    "darkgreen": (0, 0.5, 0),
+    "purple": (0.5, 0, 0.5),
+}
+
+# Target language options for ChatGPT
+LANGUAGE_OPTIONS = {
+    "简体中文": "zh-CN",
+    "繁體中文": "zh-TW",
+    "English": "en",
+    "日本語": "ja",
+    "한국어": "ko",
+    "Español": "es",
+    "Français": "fr",
+    "Deutsch": "de",
+}
+
+# Add source language options
+SOURCE_LANGUAGE_OPTIONS = {
+    "English": "en",
+    "简体中文": "zh-CN",
+    "繁體中文": "zh-TW",
+    "日本語": "ja",
+    "한국어": "ko",
+    "Español": "es",
+    "Français": "fr",
+    "Deutsch": "de",
+    "Auto": "auto",
+}
+
+def get_cache_dir():
+    """Get or create cache directory"""
+    cache_dir = Path('.cached')
+    cache_dir.mkdir(exist_ok=True)
+    return cache_dir
+
+def get_cache_key(file_content: bytes, page_num: int, translator_name: str, target_lang: str):
+    """Generate cache key for a specific page translation"""
+    # 使用文件内容的hash作为缓存key的一部分
+    file_hash = hashlib.md5(file_content).hexdigest()
+    return f"{file_hash}_page{page_num}_{translator_name}_{target_lang}.pdf"
+
+def get_cached_translation(cache_key: str) -> pymupdf.Document:
+    """Get cached translation if exists"""
+    cache_path = get_cache_dir() / cache_key
+    if cache_path.exists():
+        return pymupdf.open(str(cache_path))
+    return None
+
+def save_translation_cache(doc: pymupdf.Document, cache_key: str):
+    """Save translation to cache"""
+    cache_path = get_cache_dir() / cache_key
+    doc.save(str(cache_path))
+
+def translate_pdf_pages(doc, doc_bytes, start_page, num_pages, translator, text_color, translator_name, target_lang):
+    """Translate specific pages of a PDF document with progress and caching"""
+    WHITE = pymupdf.pdfcolor["white"]
+    rgb_color = COLOR_MAP.get(text_color.lower(), COLOR_MAP["darkred"])
+    
+    translated_pages = []
+    
+    # Create a progress bar
+    progress_bar = st.progress(0)
+    status_text = st.empty()
+    
+    for i, page_num in enumerate(range(start_page, min(start_page + num_pages, doc.page_count))):
+        status_text.text(f"Translating page {page_num + 1}...")
+        
+        # Check cache first
+        cache_key = get_cache_key(doc_bytes, page_num, translator_name, target_lang)
+        cached_doc = get_cached_translation(cache_key)
+        
+        if cached_doc is not None:
+            translated_pages.append(cached_doc)
+        else:
+            # Create a new PDF document for this page
+            new_doc = pymupdf.open()
+            new_doc.insert_pdf(doc, from_page=page_num, to_page=page_num)
+            page = new_doc[0]
+            
+            # Extract and translate text blocks
+            blocks = page.get_text("blocks", flags=pymupdf.TEXT_DEHYPHENATE)
+            
+            for block in blocks:
+                bbox = block[:4]
+                text = block[4]
+                
+                # Translate the text
+                translated = translator.translate(text)
+                
+                # Cover original text with white and add translation in color
+                page.draw_rect(bbox, color=None, fill=WHITE)
+                page.insert_htmlbox(
+                    bbox,
+                    translated,
+                    css=f"* {{font-family: sans-serif; color: rgb({int(rgb_color[0]*255)}, {int(rgb_color[1]*255)}, {int(rgb_color[2]*255)});}}"
+                )
+            
+            # Save to cache
+            save_translation_cache(new_doc, cache_key)
+            translated_pages.append(new_doc)
+        
+        # Update progress
+        progress = (i + 1) / min(num_pages, doc.page_count - start_page)
+        progress_bar.progress(progress)
+    
+    # Clear progress indicators
+    progress_bar.empty()
+    status_text.empty()
+    
+    return translated_pages
+
+def get_page_image(page, scale=2.0):
+    """Get high quality image from PDF page"""
+    # 计算缩放后的尺寸
+    zoom = scale
+    mat = pymupdf.Matrix(zoom, zoom)
+    
+    # 使用高分辨率渲染页面
+    pix = page.get_pixmap(matrix=mat, alpha=False)
+    
+    return pix
+
+def main():
+    st.set_page_config(layout="wide", page_title="PDF Translator for Human: with Local-LLM/GPT")
+    st.title("PDF Translator for Human: with Local-LLM/GPT")
+
+    # Sidebar configuration
+    with st.sidebar:
+        st.header("Settings")
+        
+        uploaded_file = st.file_uploader("Choose a PDF file", type="pdf")
+        
+        # Add source language selection
+        source_lang_name = st.selectbox(
+            "Source Language",
+            options=list(SOURCE_LANGUAGE_OPTIONS.keys()),
+            index=0  # Default to English
+        )
+        source_lang = SOURCE_LANGUAGE_OPTIONS[source_lang_name]
+        
+        translator_name = st.selectbox(
+            "Select Translator",
+            options=list(TRANSLATORS.keys()),
+            index=0
+        )
+        
+        pages_per_load = st.number_input(
+            "Pages per load",
+            min_value=1,
+            max_value=5,
+            value=DEFAULT_PAGES_PER_LOAD
+        )
+        
+        text_color = st.selectbox(
+            "Translation Color",
+            options=list(COLOR_MAP.keys()),
+            index=0
+        )
+        
+        # ChatGPT specific settings
+        if translator_name == 'chatgpt':
+            st.subheader("ChatGPT Settings")
+            target_lang = st.selectbox(
+                "Target Language",
+                options=list(LANGUAGE_OPTIONS.keys()),
+                index=0
+            )
+            api_key = st.text_input(
+                "OpenAI API Key",
+                value=os.getenv("OPENAI_API_KEY", ""),
+                type="password"
+            )
+            api_base = st.text_input(
+                "API Base URL",
+                value=os.getenv("OPENAI_API_BASE", DEFAULT_API_BASE)
+            )
+            model = st.text_input(
+                "Model Name",
+                value=os.getenv("OPENAI_MODEL", DEFAULT_MODEL)
+            )
+            
+            # Update environment variables
+            os.environ["OPENAI_API_KEY"] = api_key
+            os.environ["OPENAI_API_BASE"] = api_base
+            os.environ["OPENAI_MODEL"] = model
+            target_lang = LANGUAGE_OPTIONS[target_lang]
+        else:
+            # For Google Translator, also show target language selection
+            target_lang_name = st.selectbox(
+                "Target Language",
+                options=list(SOURCE_LANGUAGE_OPTIONS.keys())[:-1],  # Remove "Auto" option
+                index=0  # Default to first language
+            )
+            target_lang = SOURCE_LANGUAGE_OPTIONS[target_lang_name]
+
+    # Main content area
+    if uploaded_file is not None:
+        doc_bytes = uploaded_file.read()
+        doc = pymupdf.open(stream=doc_bytes)
+        
+        # Create two columns for side-by-side display
+        col1, col2 = st.columns(2)
+        
+        # Initialize session state
+        if 'current_page' not in st.session_state:
+            st.session_state.current_page = 0
+            st.session_state.translation_started = True  # 自动开始翻译
+        
+        # Display original pages immediately
+        with col1:
+            st.header("Original")
+            for page_num in range(st.session_state.current_page,
+                                min(st.session_state.current_page + pages_per_load, doc.page_count)):
+                page = doc[page_num]
+                pix = get_page_image(page)
+                st.image(pix.tobytes(), caption=f"Page {page_num + 1}", use_container_width=True)
+        
+        # Translation column
+        with col2:
+            st.header("Translated")
+            
+            # Configure translator with selected source language
+            TranslatorClass = TRANSLATORS[translator_name]
+            translator = TranslatorClass(source=source_lang, target=target_lang)
+            
+            # Translate current batch of pages
+            translated_pages = translate_pdf_pages(
+                doc,
+                doc_bytes,
+                st.session_state.current_page,
+                pages_per_load,
+                translator,
+                text_color,
+                translator_name,
+                target_lang
+            )
+            
+            # Display translated pages
+            for i, trans_doc in enumerate(translated_pages):
+                page = trans_doc[0]
+                pix = get_page_image(page)
+                st.image(pix.tobytes(), caption=f"Page {st.session_state.current_page + i + 1}", use_container_width=True)
+        
+        # Navigation buttons
+        nav_col1, nav_col2 = st.columns(2)
+        with nav_col1:
+            if st.session_state.current_page > 0:
+                if st.button("Previous Pages"):
+                    st.session_state.current_page = max(0, st.session_state.current_page - pages_per_load)
+                    st.rerun()
+        
+        with nav_col2:
+            if st.session_state.current_page + pages_per_load < doc.page_count:
+                if st.button("Next Pages"):
+                    st.session_state.current_page = min(
+                        doc.page_count - 1,
+                        st.session_state.current_page + pages_per_load
+                    )
+                    st.rerun()
+    else:
+        st.info("Please upload a PDF file to begin translation")
+
+if __name__ == "__main__":
+    main() 

pyproject.toml

@@ -30,7 +30,7 @@ description = "A flexible free and unlimited python tool to translate between di
 license = "MIT"
 authors = ["Nidhal Baccouri <[email protected]>"]
 maintainers = ["Nidhal Baccouri <[email protected]>", "Chris Trenthem <[email protected]>"]
-readme = "docs/README.rst"
+readme = "README.md"
 homepage = "https://github.com/nidhaloff/deep_translator"
 repository = "https://github.com/nidhaloff/deep_translator"
 documentation = "https://deep-translator.readthedocs.io/en/latest/"

run_translator_web.sh

@@ -0,0 +1,9 @@
+# install the project and dependencies
+
+git clone https://github.com/davideuler/pdf-translator-for-human
+cd pdf-translator-for-human
+pip install -e .
+pip install streamlit pymupdf openai 
+
+# Start the Web Application
+streamlit run pdf_translator_web.py

translator_cli.py

@@ -118,8 +118,10 @@ def main():
     export OPENAI_API_KEY=sk-proj-xxxx
     export OPENAI_API_BASE=https://api.xxxx.com/v1
     export OPENAI_API_BASE=http://localhost:8080/v1 #  for local llm api
-    python translator_cli.py --source english --translator chatgpt --target zh-CN input.pdf
+    export OPENAI_MODEL=default_model
     
+    python translator_cli.py --source english --translator chatgpt --target zh-CN input.pdf
+
     # do not keep original text as an optional layer:
     python translator_cli.py --source english --translator chatgpt --target zh-CN --no-original input.pdf