Caution
This implementation is currently not up to date and does NOT generate valid ISCCs.
A command line tool that creates ISCC Codes for digital media files based on the reference implementation.
The International Standard Content Code is a proposal for an open standard for decentralized content identification. ISCC Codes are generated algorithmically from the content itself and offer many powerful features like content similarity clustering and partial integrity checks. If you want to learn more about the ISCC please check out https://iscc.codes.
This tool offers an easy way to generate ISCC codes from the command line. It supports content extraction via Apache Tika and uses the ISCC reference implementation.
doc, docx, epub, html, odt, pdf, rtf, txt, xml, ibooks, md, xls, mobi ...
gif, jpg, png, tif, bmp, psd, eps ...
Note: EPS (postscript) support requires Ghostscript to be installed on your system and available on your PATH. (Make sure you can run gs
from your command line.)
aif, mp3, ogg, wav ...
Note: Support for the Audio-ID is experimental and not yet part of the specification
3gp, 3g2, asf, avi, flv, gif, mpg, mp4, mkv, mov, ogv, webm, wmv ...
Note: Support for the Video-ID is experimentel and not yet part of the specification
NOTE: Requires JAVA to be installed and on your path! |
---|
iscc-cli is tested on Linux, Windows, and macOS with Python 3.6/3.7/3.8.
This tool depends on tika-python. Tika is used for extracting metadata and content from media files before generating ISCC Codes. On first execution of the iscc
command line tool it will automatically download and launch the Java Tika Server in the background (this may take some time). Consecutive runs will access the existing Tika instance. You may explicitly pre-launch the Tika server with $ iscc init
The ISCC command line tool is published with the package name iscc-cli
on the Python Package Index and can be installed with pip:
$ pip3 install iscc-cli
Self-contained Windows binary executables are available for download at: https://github.com/iscc/iscc-cli/releases/
Show help overview by calling iscc
without any arguments:
$ iscc
Usage: iscc [OPTIONS] COMMAND [ARGS]...
Options:
--version Show the version and exit.
--help Show this message and exit.
Commands:
gen* Generate ISCC Code for FILE.
batch Create ISCC Codes for all files in PATH.
dump Dump Tika extraction results for PATH (file or url path).
info Show information about environment.
init Inititalize and check environment.
sim Estimate Similarity of ISCC Codes A & B.
test Test conformance with latest reference data.
web Generate ISCC Code from URL.
Get help for a specific command by entering iscc <command>
:
$ iscc gen
Usage: iscc gen [OPTIONS] FILE
Generate ISCC Code for FILE.
Options:
-g, --guess Guess title (first line of text).
-t, --title TEXT Title for Meta-ID creation.
-e, --extra TEXT Extra text for Meta-ID creation.
-v, --verbose Enables verbose mode.
-h, --help Show this message and exit.
The gen
command generates an ISCC Code for a single file:
$ iscc gen tests/image/demo.jpg
ISCC:CC1GG3hSxtbWU-CYDfTq7Qc7Fre-CDYkLqqmQJaQk-CRAPu5NwQgAhv
The gen
command is default so you can skip it and simply do $ iscc tests/demo.jpg
To get a more detailed result use the -v
(--verbose
) option:
$ iscc -v tests/image/demo.jpg
ISCC:CC1GG3hSxtbWU-CYDfTq7Qc7Fre-CDYkLqqmQJaQk-CRAPu5NwQgAhv
Norm Title: concentrated cat
Tophash: 7a8d0c513142c45f417e761355bf71f11ad61d783cd8958ffc0712d00224a4d0
Filepath: tests/image/demo.jpg
GMT: image
See iscc batch
for help on how to generate ISCC codes for multiple files at once.
The web
command allows you to create ISCC codes from URLs:
$ iscc web https://iscc.foundation/news/images/lib-arch-ottawa.jpg
ISCC:CCbUCUSqQpyJo-CYaHPGcucqwe3-CDt4nQptEGP6M-CRestDoG7xZFy
The sim
command computes estimated similarity of two ISCC Codes:
$ iscc sim CCUcKwdQc1jUM CCjMmrCsKWu1D
Estimated Similarity of Meta-ID: 78.00 % (56 of 64 bits match)
You may also compare full four-component ISCC Codes.
While this package is not built to be used as a library, some of the high level commands to generate ISCC Codes are exposed as vanilla python functions:
from iscc_cli import lib
from pprint import pprint
pprint(lib.iscc_from_url("https://iscc.foundation/news/images/lib-arch-ottawa.jpg"))
{'gmt': 'image',
'iscc': 'CCbUCUSqQpyJo-CYaHPGcucqwe3-CDt4nQptEGP6M-CRestDoG7xZFy',
'norm_title': 'library and archives canada ottawa',
'tophash': 'e264cc07209bfaecc291f97c7f8765229ce4c1d36ac6901c477e05b2422eea3e'}
Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.
Please make sure to update tests as appropriate.
You may also want join our developer chat on Telegram at https://t.me/iscc_dev.
- Update to custom mediatype detection (without Tika requirement)
- Update dependencies
- Update dependencies
- Remove support for creating ISCC codes from youtube urls
- Fixed issue with mime-type detection
- Changed wording of similarity output
- Added CSV-compatible output for batch command
- Added debug option for batch command
- Updated dependencies
- Fixed issue with tika & macOS
- Added macOS ci testing
- Updated dependencies
- Updated Content-ID-Audio for robustness against transcoding (breaking change)
- Changed similarity calculation to match with web demo
- Fixed bug in mime-type detection
- Updated dependencies
- Add support for flac and opus audio formats
- Update dependencies
- Support urls with dump command
- Updated tika 1.24 and fpcalc 1.50
- Use filename for meta-id as last resort
- Switch to signed audio fingerprint (breaking change)
- Bugfixes and stability improvements
- Support mobi7
- Support mobi print replica
- Support mobi with web command
- Add experimental support for mobi files
- Add support for XHTML
- Fix error on unsupported media types
- Add support for bmp, psd, xls, xlsx
- Add tika server live testing
- Fix error with title guess on image files
- Fix issue with APP_DIR creation
- Add experimental support for Video-ID
- Add special handling of YouTube URLs
- Add support for more Media Types (try & error)
- Add support for Python 3.8
- Remove support for Python 3.5
- Add new
test
command for confromance testing - Add support for .md (Markdown) files
- Update to ISCC v1.0.5
- Update to Apache Tika 1.23
- Fix issue with non-conformant Meta-ID
- Add support for tif files
- Add support for eps files
- Set application directory to non-roaming path
- Add new
dump
command (dumps extraction results) - Add support for iBooks files
- Fix error with tika 1.22 dependency
- Store tika server in non-volatile storage
- Expose commands as python API
- Fix title guessing bug
- Added new
web
command (creates ISCC Codes for URLs)
- Added experimental support for aif, mp3, ogg, wav
- More verbose batch output
- Fix batch output default Meta-ID
- Added support for html, odt, txt, xml, gif
- Added optional guessing of title (first line of text)
- Added new
info
command - Fixed wrong detection of identical Instance-ID
- Add
sim
command similarity comparison of ISCC Codes
- Add support for doc, docx and rtf documents
- Update to ISCC 1.0.4 (fixes whitespace bug)
- Basic ISCC Code creation
- Supported file types: jpg, png, pdf, epub
MIT © 2019-2021 Titusz Pan