Parses a subset of requirement.txt syntax

https://pip.pypa.io/en/stable/reference/requirements-file-format/

Supported:

• PEP 508 requirements
• -r
• -c
• --hash (postfix)
• -e

Unsupported:

• <path>. TBD
• <archive_url>. TBD
• Options without a requirement, such as --find-links or --index-url

Grammar as implemented:

file = (statement | empty ('#' any*)? '\n')*
empty = whitespace*
statement = constraint_include | requirements_include | editable_requirement | requirement
constraint_include = '-c' ('=' | wrappable_whitespaces) filepath
requirements_include = '-r' ('=' | wrappable_whitespaces) filepath
editable_requirement = '-e' ('=' | wrappable_whitespaces) requirement
# We check whether the line starts with a letter or a number, in that case we assume it's a
# PEP 508 requirement
# https://packaging.python.org/en/latest/specifications/name-normalization/#valid-non-normalized-names
# This does not (yet?) support plain files or urls, we use a letter or a number as first
# character to assume a PEP 508 requirement
requirement = [a-zA-Z0-9] pep508_grammar_tail wrappable_whitespaces hashes
hashes = ('--hash' ('=' | wrappable_whitespaces) [a-zA-Z0-9-_]+ ':' [a-zA-Z0-9-_] wrappable_whitespaces+)*
# This should indicate a single backslash before a newline
wrappable_whitespaces = whitespace ('\\\n' | whitespace)*

uv-requirements-txt

This crate is an internal component of uv. The Rust API exposed here is unstable and will have frequent breaking changes.

This version (0.0.48) is a component of uv 0.11.15. The source can be found here.

See uv's crate versioning policy for details on versioning.