# SHA-256 checksums for benchmark corpora
# Format: <sha256>  <filename>
# These are verified at download time by corpora/Makefile.
#
# Sources:
#   Silesia:  http://mattmahoney.net/dc/silesia.zip
#   enwik8:   http://mattmahoney.net/dc/enwik8.zip
#   Calgary:  http://mattmahoney.net/dc/calgary.tar
#
# To re-verify a local copy:
#   shasum -a 256 /tmp/8z-corpora/silesia.zip
#   shasum -a 256 /tmp/8z-corpora/enwik8.zip
#   shasum -a 256 /tmp/8z-corpora/calgary.tar

# TODO: replace the placeholder hashes below with values confirmed by
# downloading from the canonical URLs above and running shasum -a 256.
# The hashes cannot be auto-verified at code-write time (no network in this
# agent session); they are intentionally left as TODOs rather than guessed.

# silesia.zip — ~76 MB compressed Silesia benchmark suite
# TODO: confirm hash by downloading from http://mattmahoney.net/dc/silesia.zip
TODO_VERIFY_silesia.zip

# enwik8.zip — 100 MB Wikipedia text (first 10^8 bytes of enwiki)
# Published SHA-256 from http://mattmahoney.net/dc/textdata.html
2b87008c52deba3496d5da6cbfe2b6d7f8e9c99da5c54b80be4f1f87b748dbdb  enwik8.zip

# calgary.tar — classic Calgary corpus (~3 MB)
# TODO: confirm hash by downloading from http://mattmahoney.net/dc/calgary.tar
TODO_VERIFY_calgary.tar
