atrep — Python bindings

Python bindings for atrep, an extensible semantic markup language designed with literary applications in mind. The same Rust core the CLI wraps (atrep on crates.io), surfaced pythonically: strings in, strings out, exceptions instead of exit codes.

pip install atrep

Checking and the two spellings

import atrep

doc = open("apology.atd").read()

# Parse and validate; routes on the declaration.
atrep.check(doc)          # -> "document" | "dialektos"

# The two spellings: brachygraphic (sim symbols; canonical)
# and plerographic (sim names in braces). Both parse to the
# identical structure.
names = atrep.plero(doc)              # @# ... -> @{section} ...
symbols = atrep.brachy(names)         # ... and back
localized = atrep.plero(doc, lang="el")   # glossa sim names

The structural outline

o = atrep.outline(doc)
o["dialektos"]    # "litogramma"
o["blocks"]       # [{"kind", "symbol", "name", "depth",
                  #   "start", "end", "lemma", "onym",
                  #   "genoses"}, ...]
o["milestones"]   # [{"key": "steph:17a", "line": 4}, ...]
o["onyms"]        # located onym anchors
o["deixes"]       # located deixis references

Endo importers

# Foreign formats -> canonical atrep: markdown | html | rst |
# org | djot | docbook | bibtex | jats | tei | usfm | usx | osis |
# fb2 | rnc | opencorpora | proiel | conllu.
a = atrep.endo("markdown", open("notes.md").read())

# Versification milestone scheme (usfm/usx/osis only).
a = atrep.endo("usfm", open("john.usfm").read(), scheme="std")

Dialektoi

# Resolve a dialektos and get its canonical serialization.
# The std dialektoi are embedded; `dir` resolves from disk,
# `sources` from in-memory artifacts.
lekt = atrep.dialektos("koine")
lekt = atrep.dialektos("demo", sources={"demo.lektos": dia})
atrep.std_dialektoi()   # ["at-djot", "koine", ..., "litogramma"]

Errors raise atrep.AtrepError, with file:line:col locations in the message where the source provides them. atrep.version() reports the bindings/core version.

Resolution is offline: non-std dialektoi resolve from the directory of the path argument (default doc.atd, i.e. the process working directory). Remote fetching (kanonizo) and the multi-document operations (morphisms, exo rendering, collation) belong to the CLI.

License

MIT or Apache-2.0, at your option.