Merge pull request #92 from matze-dd/Issue-85

Issue 85

HISTORY.md

@@ -3,11 +3,16 @@ Work in progress
 - LaTeX macros / environments
   - builtins: added \\bibitem, \\bibliographystyle, \\begin{thebibliography}
     (issue [#80](../../issues/80))
+- yalafi.shell
+  - added option --simple-equations for simple replacements of displayed
+    equations (issue [#85](../../issues/85))
 - yalafi core
   - added 'ru' for option --lang (issue [#84](../../issues/84))
+  - added option --seqs for simple replacements of displayed equations
+    (issue [#85](../../issues/85))
   - fixed problem in tests/test_packages/test\_latex\_builtins.py
     (issue [#86](../../issues/86))
-- README.md: minor edits
+- README.md: updated
 
 Version 1.1.6 (2020/10/19)
 --------------------------

README.md

@@ -251,6 +251,10 @@ Default option values are set at the Python script beginning.
   comma-separated list.
   This is useful for check of foreign-language text, if marked accordingly.
   Internally used for detection of file inclusions on --include.
+- `--simple-equations`<br>
+  Replace a displayed equation only with a single placeholder from collection
+  'Paramaters.math\_repl\_display' in file yalafi/parameters; append trailing
+  interpunction, if present.
 - `--disable rules`<br>
   Comma-separated list of ignored LT rules, is passed as --disable to LT
   (default: 'WHITESPACE\_RULE').
@@ -1071,6 +1075,10 @@ are placed at the right side of the alignment character '\&'.
 LaTeX does not enforce that, but it is the style found in examples of the
 documentation for package amsmath.
 
+**Remark.**
+For a simplification, see option --simple-equations in section
+[Example application](#example-application).
+
 With the default entry
 ```
     EquEnv(self, 'align'),
@@ -1219,7 +1227,7 @@ The LaTeX filter can be integrated in shell scripts, compare the examples in
 ```
 python -m yalafi [--nums file] [--repl file] [--defs file] [--dcls class]
                  [--pack modules] [--extr macros] [--lang xy] [--ienc enc]
-                 [--unkn] [latexfile]
+                 [--seqs] [--unkn] [latexfile]
 ```
 Without positional argument `latexfile`, standard input is read.
 
@@ -1243,6 +1251,9 @@ Without positional argument `latexfile`, standard input is read.
   proof titles, and for handling of macros like '\"\='.
 - `--ienc enc`<br>
   As option --encoding in section [Example application](#example-application).
+- `--seqs`<br>
+  As option --simple-equations in section
+  [Example application](#example-application).
 - `--unkn`<br>
   As option --list-unknown in section
   [Example application](#example-application).
@@ -1265,6 +1276,7 @@ Invocation of `python -m yalafi ...` differs as follows from
   see [the example below](#equation-html-report).
 - Added options --dcls and --pack allow modification of predefined LaTeX
   macros and environments at Python level.
+- Added option --seqs.
 - Option --defs expects a file containing macro definitions as LaTeX code.
 - Option --ienc is also effective for file from --defs.
 - Option --char (position tracking for single characters) is always activated.

tests/test_display.py

@@ -186,3 +186,22 @@ def test_9():
     plain, pos = utils.get_txt_pos(toks)
     assert plain_9 == plain
 
+#   simplified equation parsing
+#
+latex_10 = r"""
+\usepackage{amsmath}
+\begin{align}
+    a &= b.
+\end{align}
+"""
+plain_10 = r"""
+  W-W-W.
+"""
+def test_10():
+    parms = parameters.Parameters()
+    parms.math_displayed_simple = True
+    p = parser.Parser(parms)
+    toks = p.parse(latex_10)
+    plain, pos = utils.get_txt_pos(toks)
+    assert plain_10 == plain
+

yalafi/mathparser.py

@@ -61,7 +61,7 @@ def __init__(self, parser):
 
     def expand_display_math(self, buf, tok, env):
         buf.next()
-        start = tok.pos
+        start_simple = start = tok.pos
         first_section = True
         next_repl = True
         out = [defs.ActionToken(start),
@@ -93,6 +93,15 @@ def expand_display_math(self, buf, tok, env):
             else:
                 out = [defs.ActionToken(out[-1].pos)]
         else:
+            if self.parser.parms.math_displayed_simple:
+                txt = self.parser.get_text_direct(out).strip()
+                out = [defs.ActionToken(start_simple),
+                        defs.SpaceToken(start_simple, '  ', pos_fix=True),
+                        defs.TextToken(start_simple,
+                        self.parser.parms.math_repl_display[0], pos_fix=True)]
+                if txt and txt[-1] in self.parser.parms.math_punctuation:
+                    out.append(defs.TextToken(out[-1].pos, txt[-1],
+                                                        pos_fix=True))
             out.append(defs.ActionToken(out[-1].pos))
         return out
 

yalafi/parameters.py

@@ -329,6 +329,10 @@ def init_collections(self):
     #
     def init_math_collections(self):
 
+        #   simple replacement for displayed equations?
+        #
+        self.math_displayed_simple = False
+
         #   things to be ignored in math mode
         #   - some entries are redundant, if macros are known from text mode
         #     and expand to 'nothing'

yalafi/shell/proofreader.py

@@ -62,6 +62,7 @@ def run_proofreader_options(tex, language, disable, enable,
     t2t_options = tex2txt.Options(char=True, repl=cmdline.replace,
                             defs=cmdline.define, lang=language[:2],
                             extr=cmdline.extract, unkn=cmdline.list_unknown,
+                            seqs=cmdline.simple_equations,
                             dcls=cmdline.documentclass, pack=cmdline.packages)
 
     if cmdline.plain_input:

yalafi/shell/shell.py

@@ -138,6 +138,7 @@
 parser.add_argument('--skip')
 parser.add_argument('--plain-input', action='store_true')
 parser.add_argument('--list-unknown', action='store_true')
+parser.add_argument('--simple-equations', action='store_true')
 parser.add_argument('--language', default=default_option_language)
 parser.add_argument('--encoding', default=default_option_encoding)
 parser.add_argument('--replace')

yalafi/tex2txt.py

@@ -44,6 +44,8 @@ def read(file):
         extr = ['\\' + s for s in opts.extr.split(',')]
     else:
         extr = []
+    if opts.seqs:
+        parms.math_displayed_simple = True
     p = parser.Parser(parms, packages, read_macros=read)
     toks = p.parse(latex, extract=extr)
     txt, pos = utils.get_txt_pos(toks)
@@ -199,6 +201,7 @@ def __init__(self,
             pack=None,      # import modules for \usepackage
             extr=None,      # or string: comma-separated macro list
             lang=None,      # or set to language code
+            seqs=False,     # True: simple replacements for displayed equations
             unkn=False):    # True: print unknowns
         self.ienc = ienc
         self.repl = repl
@@ -211,6 +214,7 @@ def __init__(self,
         self.pack = pack
         self.extr = extr
         self.lang = lang
+        self.seqs = seqs
         self.unkn = unkn
 
 #   function to be called for stand-alone script
@@ -227,6 +231,7 @@ def main():
     parser.add_argument('--extr')
     parser.add_argument('--lang')
     parser.add_argument('--ienc')
+    parser.add_argument('--seqs', action='store_true')
     parser.add_argument('--unkn', action='store_true')
     cmdline = parser.parse_args()
 
@@ -242,6 +247,7 @@ def main():
                 pack=cmdline.pack,
                 extr=cmdline.extr,
                 lang=cmdline.lang,
+                seqs=cmdline.seqs,
                 unkn=cmdline.unkn)
 
     if cmdline.file: