“Many industrial software or legacy system manuals are in the compiled HTML Help (.chm) format, which is a closed format that modern browsers cannot translate directly.”

Root Cause Analysis

Automated Decompiling

After uploading a CHM file, the Shangyi AI backend will invoke `hh.exe` or open-source libraries to decompile it into the original set of HTML files (topic files) and index files (HHK/HHC).

Link Rewriting and Localization

The system translates index files and table of contents files concurrently with HTML content. More importantly, it corrects internal relative links within the HTML to ensure accurate page navigation after translation.

Encoding Conversion (GB2312/UTF-8)

Legacy CHM files commonly use ANSI encoding, such as GB2312. During processing, the AI automatically converts them to UTF-8 for translation. When recompiling, you can choose to keep UTF-8 or revert to the system default encoding of the target language to prevent garbled characters.

Final Solution Summary

Full localization of legacy software help systems can be achieved without the original source code.