spaCy/website/create_code_samples

77 lines
2.4 KiB
Python
Executable File

#!/usr/bin/env python
from __future__ import unicode_literals
import os
import ast
import io
import re
import plac
# cgi.escape is deprecated since py32
try:
from html import escape
except ImportError:
from cgi import escape
# e.g. python website/create_code_samples tests/website/ website/src/
def main(src_dirname, dst_dirname):
prefix = "test_"
for filename in os.listdir(src_dirname):
if not filename.startswith('test_'):
continue
if not filename.endswith('.py'):
continue
# Remove test_ prefix and .py suffix
name = filename[5:-3]
with io.open(os.path.join(src_dirname, filename), 'r', encoding='utf8') as file_:
source = file_.readlines()
tree = ast.parse("".join(source))
for root in tree.body:
if isinstance(root, ast.FunctionDef) and root.name.startswith(prefix):
# only ast.expr and ast.stmt have line numbers, see:
# https://docs.python.org/2/library/ast.html#ast.AST.lineno
line_numbers = []
for node in ast.walk(root):
if hasattr(node, "lineno"):
line_numbers.append(node.lineno)
body = source[min(line_numbers)-1:max(line_numbers)]
while not body[0][0].isspace():
body = body[1:]
# make sure we are inside an indented function body
assert all([l[0].isspace() for l in body])
offset = 0
for line in body:
match = re.search(r"[^\s]", line)
if match:
offset = match.start(0)
break
# remove indentation
assert offset > 0
for i in range(len(body)):
body[i] = body[i][offset:] if len(body[i]) > offset else "\n"
# make sure empty lines contain a newline
assert all([l[-1] == "\n" for l in body])
code_filename = "%s.%s" % (name, root.name[len(prefix):])
with io.open(os.path.join(dst_dirname, code_filename),
"w", encoding='utf8') as f:
f.write(escape("".join(body)))
if __name__ == '__main__':
plac.call(main)