Identify and Translate PDF

Warning

After version 0.3.1 the output has been updated to logging, which by default only outputs Warning and above. If you want to see the processing, set the logging level to INFO:

import logging
httpx_logger = logging.getLogger("httpx")
httpx_logger.setLevel(logging.WARNING)
logging.basicConfig(level=logging.INFO)

For demonstration purposes, the following code examples all set the logging level to INFO.

Client.pdf_translate

Caution

Please note that this interface is not officially supported and is not guaranteed to be available.

Translate one or more PDF files into text files in the specified language.

Parameters

Parameter	Type	Required	Default	Description
pdf_file	str or list	Yes	-	Path to the PDF file or list of PDF file paths
output_path	str	No	"./Output"	Path to the output folder
convert	bool	No	False	Whether to convert [ to $, and [[ to $$
language	str	No	"zh"	Target language, supported languages: "en", "zh", "ja", "fr", "ru", "pt", "es", "de", "ko", "ar"
model	str	No	"deepseek"	Translation model, supported models: "deepseek", "glm4"

Return Values

Returns a tuple (list1, list2, status) containing three elements, in the same order as the input files:

1. list1 (list): List of successfully translated files

   • Elements are translated text and text location (strings)
   • Empty string if processing failed

2. list2 (list): List of files that failed processing

   • Elements are dictionaries containing two keys:
     • 'error': Error message (string)
     • 'path': Path of the file that failed processing (string)
   • Both keys' values are empty strings if processing succeeded

3. status (bool): Processing status

   • True: At least one file failed processing
   • False: All files processed successfully

Notes

• The lengths of list1 and list2 are the same.
• If the API key does not have translation permissions, a RuntimeError exception will be thrown.

Warning

The return value of this function's list1 is different from other functions; see below for details.

Detailed Explanation of Return Values

The returned list1 contains two sublists:

1. text["texts"] (list): List of translated texts

   • Elements are translated texts (strings)
   • Empty string indicates that the current text block was not translated (e.g., it is table text)

2. text["location"] (list): List of text location information

   • Elements are text location information (strings)
   • Corresponds to each translated text in text["texts"], indicating its position in the original PDF

Example

Warning

Please ensure you have configured your API key in environment variables as described in the Initialization section.

from pdfdeal import Doc2X

Client = Doc2X()
translate, fail, flag = Client.pdf_translate(
    pdf_file="tests/pdf/sample.pdf", language="zh", model="deepseek"
)
for text in translate:
    print(text["texts"])
    print(text["location"])
print(fail)
print(flag)

Expected output, where dark areas represent printed variable values:

Processing file: 6%    -- uuid: 655947fa-277c-4f05-8edc-b92f0eca3a63
TRANSLATE Progress: 1/1 files successfully processed.
['## 测试', '\n\n## 测试']
[{'raw_text': '## Test', 'page_idx': 0, 'page_width': 2040, 'page_height': 1148, 'x': 867, 'y': 418}, {'raw_text': '\n\n## 测试', 'page_idx': 1, 'page_width': 2040, 'page_height': 1148, 'x': 869, 'y': 412}]
[{'error': '', 'path': ''}]
False