From c2c46dbc31d6f1d57b54dd07d2398c73c5bef8e5 Mon Sep 17 00:00:00 2001
From: Tim Stadtmann <tim.stadtmann@rwth-aachen.de>
Date: Fri, 2 Dec 2016 18:43:58 +0100
Subject: [PATCH] Add script for preprocessing before doc-build
The MATLAB-extension for Sphinx runs into an infinite loop if a class
contains an enumeration definition. The implemented python parser script
comments out all enumerations (a reverse function is also implemented).
In a future release, a release-script will automate the calls of
preprocess(), make html and postprocess().
---
docs/process.py | 90 +++++++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 90 insertions(+)
create mode 100644 docs/process.py
diff --git a/docs/process.py b/docs/process.py
new file mode 100644
index 0000000..1eb28d8
--- /dev/null
+++ b/docs/process.py
@@ -0,0 +1,90 @@
+#!/usr/bin/python3
+
+import os
+from os import close, remove
+from tempfile import mkstemp
+from shutil import move
+
+def preprocess():
+ """Preprocesses data for building documentation
+
+ The MATLAB Sphinx extension takes veeery (or even infinitely?) long
+ to build the documentation if a class contains an enumeration.
+ """
+
+ code_dir = os.path.abspath('../source')
+ files = _matlab_files(os.listdir(code_dir))
+
+ n_files = len(files)
+ for n, file_name in enumerate(files):
+ print("Preprocessing file {0} [{1}|{2}]".format(file_name, n+1, n_files))
+ try:
+ fh, abs_path = mkstemp()
+ with open(code_dir + "/" + file_name, 'r') as f:
+ with open(abs_path, 'w') as ftemp:
+ content = f.readlines()
+ if content[0] == '%TEMP-FILE FOR BUILDING DOCUMENTATION\n':
+ raise RuntimeError('Found already preprocessed file.')
+
+ try:
+ for i, line in enumerate(content): # Worst runtime ever
+ if _ignore_line(line) is False and 'enumeration' in line:
+ ftemp.write('%TEMP-FILE FOR BUILDING DOCUMENTATION\n')
+ _write_lines(content[:i], ftemp)
+ for j, enum_line in enumerate(content[i:]):
+ ftemp.write(" %"+enum_line)
+ if " end" in enum_line or "end\n" == enum_line:
+ _write_lines(content[i+1+j:], ftemp)
+ close(fh)
+ remove(code_dir + "/" + file_name)
+ move(abs_path, code_dir + "/" + file_name)
+ # At this point, entire file has been processed: continue outer loop
+ # Hackily implemented by raising an exception
+ raise StopIteration
+ except StopIteration:
+ continue
+ except IOError:
+ pass
+
+def postprocess():
+ code_dir = os.path.abspath('../source')
+ files = _matlab_files(os.listdir(code_dir))
+
+ n_files = len(files)
+ for n, file_name in enumerate(files):
+ print("Postprocessing file {0} [{1}|{2}]".format(file_name, n+1, n_files))
+ try:
+ fh, abs_path = mkstemp()
+ with open(code_dir + "/" + file_name, 'r') as f:
+ with open(abs_path, 'w') as ftemp:
+ content = f.readlines()
+ if content[0] != '%TEMP-FILE FOR BUILDING DOCUMENTATION\n':
+ continue
+ try:
+ for i, line in enumerate(content): # Worst runtime ever
+ if 'enumeration' in line:
+ _write_lines(content[1:i], ftemp)
+ for j, enum_line in enumerate(content[i:]):
+ ftemp.write(enum_line.replace(' %', '', 1))
+ if " end" in enum_line or "end\n" == enum_line:
+ _write_lines(content[i+1+j:], ftemp)
+ close(fh)
+ remove(code_dir + "/" + file_name)
+ move(abs_path, code_dir + "/" + file_name)
+ # At this point, entire file has been processed: continue outer loop
+ # Hackily implemented by raising an exception
+ raise StopIteration
+ except StopIteration:
+ continue
+ except IOError:
+ pass
+
+def _ignore_line(line) -> bool:
+ return (line.startswith('%') or line == '\n' or line == '')
+
+def _write_lines(lines, f):
+ for line in lines:
+ f.write(line)
+
+def _matlab_files(files) -> list:
+ return [f for f in files if (f.endswith(".m") and "_" not in f and "proto" not in f)]
--
GitLab