These directions are for Linux

Below:
  "GD"="Google Docs"
  "PB"="Pressbooks"

The goal of this workflow is to make an accessibility ["a11y"] report about the
GD as it will be transferred to PB or some other platform.

Download GD file by
  File > Download > Web Page (.html, zipped)
let's say into an otherwise empty directory "work," where we will assume you
are now working.

Do:
% unzip *.zip
% mv *.html book.html
% /path/OOshow_colors book.html

Look in book.html for text with the listed classes that is bad -- typically,
various not-quite-blacks are fine, and often blue is for links, but make sure
there aren't significant bits of text with another color that is used to convey
meaning.

a11y report: indicate whether any bad color<->meaning text is found in
             book.html and, if so, where

Do:
% /path/OOprep -v book.html
% /path/OOoutliner -n -t tidy_book.html | less

There should be no improperly nested headers (OOoutline will warn about them)
All h1 tags should have the form "<h1>Chapter <number>: <chapter title></h1>"
There should be no <img> tags in headers.
Probably no &nbsp; in headers.

a11y report: indicate whether any improperly nested headers are found, clean
             up other issues

Handling figures:
The goal is to get figure notations into a specific format, so that they can
be automatically extracted for a11y investigation.

First, let's make sure that just the word "Figure" is correctly capitalized
and not abbreviated, by doing:

% /path/OOcapitalize_figures -v tidy_book.html

Do some by-hand fixing of general formatting issues, first in emacs
in capfig_tidy_book.html: replace-regexp
<p id="[^"]*>
with
<p>

replace-string
<p>&nbsp;
with
<p>

replace-string
igure&nbsp
with
igure 
(note space at the end of that line)

search for the string image description, make sure they are all in reasonable places

Do some by-hand fixing of <img> tags which are not well formatted.  Can find
them by using
% /path/OOfind_img_tags -v -b -n -s capfig_tidy_book.html
which shows the lines with badly located <img> tags, along with their line
numbers.  Variants can be useful:
% /path/OOfind_img_tags -v -b -n capfig_tidy_book.html
shows just the line numbers of those badly located <img> tags
% /path/OOfind_img_tags -v -c -n -s capfig_tidy_book.html
shows the lines where there is an <img> tag not followed by a caption line
and the caption line not following an <img> tag line
% /path/OOfind_img_tags -v -c -n capfig_tidy_book.html
shows just the lines numbers of an <img> tag not followed by a caption line
and of a caption line not following an <img> tag line
% /path/OOfind_img_tags -v -m -n -s capfig_tidy_book.html
shows sequential blocks of lines all with <img> tags
% /path/OOfind_img_tags -v -m -n capfig_tidy_book.html
shows just the line numbers of those sequential blocks
% /path/OOfind_img_tags -v -a -n -s capfig_tidy_book.html
shows all lines with <img> tags
% /path/OOfind_img_tags -v -a -n capfig_tidy_book.html
shows just the line numbers of all lines with <img> tags

clean up <img> tag and caption lines found above; in particular,
make sure the first occurrance of a caption line is immediately
following the appropriate image.  If not, can insert a &npbs; between
the <p> and the Figure on that caption-looking line to get later programs
not to treat it as a caption line.

also for images:
make all multi-img figures which have lines like
<p><img ..></p>
<p><img ..></p>
<p>Figure x.y ...</p>
into
<p><img ..><img ..></pb>
<p>Figure x.y ...</p>

% /path/OOfig_finder -v -n -c capfig_tidy_book.html -o <outputfilename>

go through <outputfilename>.csv in LibreOffice and see if there are unknown
figure types or missing figures, maybe tinker with capfig_tiday_book.html to
fix them, run OOfig_finder again

working here... 

Go through the

search for "<table>" in tidy_book.html
fix epigraphs
leave 1-col-1-row tables around LandA, a few other recognized structures
remove other (superfluous) 1-col-1-row tables, leave note in doc if unsure

to find strangely located <table>s, do: (this is ok for <tables in LandAs and int devpre2ad text which numbers them separately and keeps them as "Table X.Y":

fh = open("tidy_book.html","r")
tb=fh.read()
tbl=tb.split("\n")
import re
patt=re.compile("<table")
patT=re.compile("<p>Table")
patla=re.compile("<h3>Licenses and Attributions")
patH=re.compile("<h[1-5]")
Tlnl = []
tlnl = []
lalnl = []
ln = 0
lastHln = -1
lastHl = []
for l in tbl:
  if patH.match(l):
    lastHln = ln
  lastHl.append(lastHln)
  ln += 1
  if patT.match(l):
    Tlnl.append(ln)
  if patt.match(l):
    tlnl.append(ln)
  elif patla.match(l):
    lalnl.append(ln)

lawotl = [n for n in lalnl if not (n+1) in tlnl]
twolal = [n for n in tlnl if not (n-1) in lalnl]
twoTorlal = [n for n in tlnl if ((not (n-1) in lalnl) and (not (n-1) in Tlnl))]
if lawotl:
  print("Found '<h3>Licenses and Attributions...'s NOT followed by '<table's at line numbers: "+" ".join([str(x) for x in lawotl]))

if twolal:
  print("Found '<table's NOT following '<h3>Licenses and Attributions...'s at line numbers: "+" ".join([str(x) for x in twolal]))

if twoTorlal:
  print(f"Found {len(twoTorlal)} '<table's NOT following either '<h3>Licenses and Attributions...' or '<p>Table' at line numbers: "+" ".join([str(x) for x in twoTorlal]))

for x in twoTorlal:
  print(f"{x} in section {tbl[lastHl[x]]}")


to find strangely located <table>s, do: (this is ok for <tables in LandAs and int socialproblems which wraps talbes around "<h3>Going Deeper"s

fh = open("tidy_book1.html","r")
tb=fh.read()
tbl=tb.split("\n")
import re
patt=re.compile("<table")
patGD=re.compile("<h3>Going Deeper")
patla=re.compile("<h3>Licenses and Attributions")
patH=re.compile("<h[1-5]")
GDlnl = []
tlnl = []
lalnl = []
ln = 0
lastHln = -1
lastHl = []
for l in tbl:
  if patH.match(l):
    lastHln = ln
  lastHl.append(lastHln)
  ln += 1
  if patGD.match(l):
    GDlnl.append(ln)
  if patt.match(l):
    tlnl.append(ln)
  elif patla.match(l):
    lalnl.append(ln)

lawotl = [n for n in lalnl if not (n+1) in tlnl]
twolal = [n for n in tlnl if not (n-1) in lalnl]
twoGDorlal = [n for n in tlnl if ((not (n-1) in lalnl) and (not (n+3) in GDlnl))]
if lawotl:
  print(f"Found {len(lawotl)}'<h3>Licenses and Attributions...'s NOT followed by '<table's at line numbers: "+" ".join([str(x) for x in lawotl]))

if twolal:
  print(f"Found {len(twolal)}}'<table's NOT following '<h3>Licenses and Attributions...'s at line numbers: "+" ".join([str(x) for x in twolal]))

if twoGDorlal:
  print(f"Found {len(twoGDorlal)} '<table's either NOT following '<h3>Licenses and Attributions...' or not followed by '<h3>Going Deeper' at line numbers: "+" ".join([str(x) for x in twoGDorlal]))

for x in twoGDorlal:
  print(f"{x} in section {tbl[lastHl[x]]}")