This repository has moved to https://codeberg.org/hothothotdog/search_extract
This GitHub copy is a frozen archive. New commits, issues, and pull requests should go to Codeberg.
A Windows command-line tool that recursively scans a directory tree for ZIP and ARC archives, extracts them using the best available extractor, sorts the output by content type, and produces a full extraction report — all inside a live curses terminal dashboard.
It is designed to handle archives from the DOS/early PC era, including formats and compression methods that modern tools struggle with, such as PKzip Reduce compression which was a precursor to the more modern zip and the original ARC archives created in the 80's. This script will work equally well on a single archive.
After obtaining a dump of 1980-1990 archives I originally looked for a way to mass extract and document them but, found none truly suitable.
Inside all these archives there were multiple disk image types, nested archives within archives, ancient .arc and .zip files that were a pain to work on in en-mass with DOSBox. They all needed sorting too and many had CRC errors that i wanted to deal with quickly, log and put aside separately.
| Python 3.8+ | Runtime |
pip install windows-curses
| 7-Zip (7z on PATH) | Primary ZIP extractor (Deflate, Deflate64, LZMA, etc.) |
| File | Purpose |
|---|---|
msdos.exe |
DOS emulator for running legacy tools |
PKUNZIP.EXE |
Legacy extractor for Reduce-compressed ZIPs (methods 2–5) |
PKUNPAK.EXE |
Legacy extractor for ARC archives |
msdos.exe, PKUNZIP.EXE, and PKUNPAK.EXE are checked at startup. The script exits with a clear error message if any are missing before curses takes over the terminal.
Each executable requires a few extraction steps from its source archive. All extraction steps below use 7-Zip unless otherwise noted.
Alternatively, you can skip the extraction steps and download the prerequisites.7z from the releases page.
Provides: PKUNPAK.EXE
| Download file | pkarc-v3-61_35dd_ima_en.zip |
| Topic | http://www.win3x.org/win3board/viewtopic.php?t=28463&view=min |
| Download | http://www.win3x.org/win3board/ext/win3x/download.php?id=4316 |
Extraction steps:
- Extract
pkarc-v3-61_35dd_ima_en.zipwith 7-Zip - Extract
PK361.EXEfromdisk.imawith 7-Zip - Extract
PK361.EXEvia terminal:msdos.exe PK361.EXE - Copy
PKUNPAK.EXEto the script folder
Provides: PKUNZIP.EXE
| Download file | pkzip_204g_dd35_dos_en.zip |
| Topic | http://www.win3x.org/win3board/viewtopic.php?t=4336&hilit=pkunzip |
| Download | http://www.win3x.org/win3board/ext/win3x/download.php?id=4168 |
Extraction steps:
- Extract
pkzip_204g_dd35_dos_en.zipwith 7-Zip - Extract
PKUNZIP.EXEfromdisk.imawith 7-Zip - Copy
PKUNZIP.EXEto the script folder
Provides: msdos.exe
There are many versions to choose from inside the download. I went with the simplest x86 for speed.
| Download file | msdos.7z |
| Topic | https://takeda-toshiya.my.coocan.jp/msdos/index.html |
| Download | https://takeda-toshiya.my.coocan.jp/msdos/msdos.7z |
Extraction steps:
- Extract
msdos.7zwith 7-Zip - Navigate to
\msdos\binary\i86_x86\ - Copy
msdos.exeto the script folder
To view comments as intended from archives back in the day this does an excellent job.
It's a free, small and fast text editor that supports ASCII art with the extension .nfo
| Download file | nfopad181.exe |
| Main page | https://truehumandesign.se/s_nfopad.php |
| Download | https://truehumandesign.se/dl.php?file=nfopad181.exe |
# Process every ZIP and ARC found under the current directory
python search_extract.py
# log file captures both panels in real time with
python search_extract.py --log <filename>
# Only process archives that contain disk image files (see Target Extensions below)
python search_extract.py --ext
# Search for target extensions and log live output to file
python search_extract.py --ext --log live.txt
When passed, only archives whose contents match the target extension list or the wildcard pattern *.?@? are extracted. Archives with no matching content are counted as Skipped and left untouched. Without the flag every archive is processed unconditionally and the Skipped counter is hidden.
The following file extensions are recognised as disk image files when --ext is active:
.ima .flp .dd .raw .td0 .fdd .vfd .sdi .cp2 .dmg .pdi .ana .imd .ddi .dsk .img .sqz
The wildcard pattern *.?@? additionally catches non-standard disk image naming conventions.
## Edit the script to add or remove extensions
Find `def process_archives`
The snippet below shows what needs editing.
# The image extensions to search for INSIDE the archives
target_exts = ('.ima', '.flp', '.dd', '.raw', '.td0', '.fdd',
'.vfd', '.sdi', '.cp2', '.dmg', '.pdi', '.ana',
'.imd', '.ddi', '.dsk', '.img', '.sqz')
wildcard_pattern = "*.?@?"
All output lands alongside the script. Source directories are cleaned up (deleted if empty) after processing completes.
(working directory)/
│
├── _sorted_zip/ ← extracted ZIP contents
│ ├── .img/ ← single-extension match
│ ├── .ima/
│ ├── mixed/ ← multiple extensions in one archive
│ ├── _wildcard/ ← wildcard-only match
│ └── CRC-Errors/ ← failed ZIPs, sorted by extension
│ ├── .img/
│ └── __Reduce/ ← Reduce ZIPs that pkunzip failed on
│
├── _sorted_arc/ ← extracted ARC contents
│ ├── Arc/ ← successfully extracted ARCs
│ └── CRC-Errors/ ← ARCs that failed CRC check
│
├── YYYY-MM-DD_HH-MM-SS_<folder>_Report.txt ← extraction report
└── error_log.txt ← Python / 7-Zip errors
Every processed archive produces a __<filename>__metadata.nfo file. Metadata is written in raw binary mode to preserve original encoding.
Source ZIP: disk001.zip
Disk Label: DISK 1
Archive Comment: <raw bytes of original comment>
Source ARC: 0UTILS.ARC
Archive Comment: <raw bytes of original comment>
After a ZIP is extracted, its output folder is recursively scanned for further ZIPs up to a depth of 10. Metadata is saved for nested archives.
Source directories are only deleted if they are empty after processing — any non-archive files present are left in place and the directory is kept.
The script runs inside a full-screen curses TUI.
### Stat descriptions
| Stat | Description |
|---|---|
| **Archives** | Total archives extracted so far (ZIPs + ARCs) |
| **Successful** | Extractions that completed without error |
| **Queued** | Remaining archives still to be processed — counts down to zero in real time |
| **Legacy** | ZIPs using legacy Reduce compression, routed through pkunzip |
| **Zip Errors** | ZIPs that failed CRC verification |
| **Files** | Live count of all files present in `_sorted_zip` and `_sorted_arc` |
| **ARC** | ARC archives successfully extracted |
| **ARC Errors** | ARCs that failed CRC check |
| **Skipped** | *(--ext mode only)* Archives with no matching content, left untouched |
Recent Activity (left) — one line per archive processed, colour-coded: green for success, yellow for Reduce, red for CRC errors.
Live Extraction (right) — Terminal output streamed directly as each file is written to disk.
A timestamped report is written to the working directory on completion and opened automatically:
YYYY-MM-DD_HH-MM-SS__Report.txt
The report is structured as:
EXTRACTION SUMMARY REPORT
============================================================
Date/Time : 2025-03-14 09:41:22
Duration : 0:04:37.112894
Total Processed : 380
Successful : 371
Reduce : 5
CRC Errors : 4
Arc : 12
Arc CRC Errors : 1
Target Files Found : ---
------------------------------------------------------------
============================================================
ARC
============================================================
SUCCESS: Extracted DISKSET1.ARC
Location: _sorted_arc\Arc\DISKSET1
...
============================================================
FAILED
============================================================
CRC Error: CORRUPT.ZIP
Location: _sorted_zip\CRC-Errors\.img\CORRUPT
...
============================================================
SUCCESS
============================================================
SUCCESS: Extracted and Moved DISK001.ZIP
Location: _sorted_zip\.img\DISK001 --->
[OK] DISK.IMG
...
error_log.txt captures errors that don't fit the known categories:
- Unknown 7-Zip errors
- Python exceptions during processing
- Nested ZIP extraction failures (with depth level)
- ARC processing exceptions
The ZIP and its metadata are moved to _sorted_zip/CRC-Errors/<ext>/.
The ZIP and its metadata are moved to _sorted_zip/CRC-Errors/__Reduce/.
The ARC, its metadata and any files that could be extracted are moved to _sorted_arc/CRC-Errors/.
Output is written to error_log.txt.
Checks are made before a file not found error is encountered.
Timing problems with NTFS are worked around with retrys.
Read-only is cleared before retrying.
Files that got removed whilst the script is running are skipped and logged silently.
Catches any unexpected Python error.