From 2879df7c14c4ff8a9a54d03835b4f678c68d67ad Mon Sep 17 00:00:00 2001
From: Michal Hoftich <michal.h21@gmail.com>
Date: Wed, 25 Sep 2024 16:24:09 +0200
Subject: [PATCH] Added make4ht-char-def

---
 .gitignore                       |  1 +
 CHANGELOG.md                     |  7 ++++++
 Makefile                         | 13 ++++++----
 domfilters/make4ht-sectionid.lua |  7 +++---
 tools/make_chardata.lua          | 43 ++++++++++++++++++++++++++++++++
 5 files changed, 63 insertions(+), 8 deletions(-)
 create mode 100644 tools/make_chardata.lua

diff --git a/.gitignore b/.gitignore
index 1de04a7..fbd6c5d 100644
--- a/.gitignore
+++ b/.gitignore
@@ -21,3 +21,4 @@ tags
 *.css
 readme.tex
 changelog.tex
+make4ht-char-def.lua
diff --git a/CHANGELOG.md b/CHANGELOG.md
index 0ebf5e8..e481dfe 100755
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -1,5 +1,12 @@
 # Changelog
 
+- 2024/09/25 
+
+  - added `make4ht-char-def` library to remove dependency on `char-def` from
+    ConTeXt. It is used by the `sectionid` DOM filter. The library is
+    automatically created from UnicodeData.txt by `tools/make_chardata.lua`.
+    https://tex.stackexchange.com/q/727202/2891
+
 - 2024/09/18 
 
   - print error messages from the `log` file.
diff --git a/Makefile b/Makefile
index 87cde97..0a8de2b 100644
--- a/Makefile
+++ b/Makefile
@@ -54,9 +54,9 @@ ifeq ($(strip $(shell git rev-parse --is-inside-work-tree 2>/dev/null)),true)
 	git fetch --tags
 endif 
 
-doc: $(doc_file) readme.tex 
+doc: chardef $(doc_file) readme.tex 
 
-htmldoc: ${htmldoc}
+htmldoc: chardef ${htmldoc}
  
 make4ht-doc.pdf: $(doc_sources)
 	latexmk -pdf -pdflatex='lualatex "\def\version{${VERSION}}\def\gitdate{${DATE}}\input{%S}"' make4ht-doc.tex
@@ -70,7 +70,7 @@ readme.tex: README.md
 changelog.tex: CHANGELOG.md
 	pandoc -f markdown+definition_lists -t LaTeX CHANGELOG.md > changelog.tex
 
-build: doc $(lua_content) $(filters) $(domfilters)
+build: chardef doc $(lua_content) $(filters) $(domfilters)
 	@rm -rf build
 	@mkdir -p $(BUILD_MAKE4HT)
 	@mkdir -p $(BUILD_MAKE4HT)/filters
@@ -86,10 +86,10 @@ build: doc $(lua_content) $(filters) $(domfilters)
 	@cp README.md $(BUILD_MAKE4HT)/README
 	@cd $(BUILD_DIR) && zip -r make4ht.zip make4ht
 
-install: doc $(lua_content) $(filters) $(domfilters) justinstall 
+install: chardef doc $(lua_content) $(filters) $(domfilters) justinstall 
 	cp  $(doc_file) $(MANUAL_DIR)
 
-justinstall:
+justinstall: chardef
 	mkdir -p $(INSTALL_DIR)
 	mkdir -p $(MANUAL_DIR)
 	mkdir -p $(FILTERS_DIR)
@@ -106,6 +106,9 @@ justinstall:
 	echo $(wildcard $(EXECUTABLE))
 	$(INSTALL_COMMAND)
 
+chardef:
+	texlua tools/make_chardata.lua > make4ht-char-def.lua
+
 version:
 	echo $(VERSION), $(DATE)
 
diff --git a/domfilters/make4ht-sectionid.lua b/domfilters/make4ht-sectionid.lua
index c8e9892..25f54e6 100644
--- a/domfilters/make4ht-sectionid.lua
+++ b/domfilters/make4ht-sectionid.lua
@@ -2,13 +2,14 @@ local mkutils   = require "mkutils"
 local log = logging.new("tocid")
 -- Unicode data distributed with ConTeXt
 -- defines "characters" table
-if not mkutils.isModuleAvailable("char-def") then
+if not mkutils.isModuleAvailable("make4ht-char-def") then
   log:warning("char-def module not found")
   log:warning("cannot fix section id's")
   return function(dom) return dom end
 end
-require "char-def"
-local chardata = characters.data or {}
+
+local chardata = require "make4ht-char-def"
+
 
 
 local toc = nil
diff --git a/tools/make_chardata.lua b/tools/make_chardata.lua
new file mode 100644
index 0000000..b4d4fc3
--- /dev/null
+++ b/tools/make_chardata.lua
@@ -0,0 +1,43 @@
+kpse.set_program_name "luatex"
+-- create Lua module from UnicodeData
+-- we need mapping to lower case letters and decomposed base letters for accented characters
+local unicode_data = kpse.find_file("UnicodeData.txt")
+local chardata = {}
+for line in io.lines(unicode_data) do
+  local record = line:explode(";")
+  local char = tonumber(record[1], 16)
+  local category  = string.lower(record[3])
+  if category:match("^l") or category == "zs" then
+    -- the decomposed field contains charcode for the base letter and accent
+    -- we care only about the base letter
+    local decomposed = record[6]:match("([%x]+)")
+    decomposed = decomposed and tonumber(decomposed, 16)
+    -- the lowercase letter is the last field
+    local lower = record[#record - 1]
+    lower = lower and tonumber(lower, 16) or nil
+    chardata[#chardata+1] = {
+      char   = char,
+      shcode = decomposed,
+      lccode = lower,
+      category = category
+    }
+  end
+end
+
+print "return {"
+local function add(fields, caption, value)
+  if value then
+    fields[#fields+1] = string.format("%s=%s", caption, value)
+  end
+end
+
+for _, data in ipairs(chardata) do
+  local fields = {}
+  -- we need to add qotes to force string
+  add(fields, "category", string.format('"%s"', data.category))
+  add(fields, "lccode", data.lccode)
+  add(fields, "shcode", data.shcode)
+  print(string.format("[%s] = {%s},", data.char, table.concat(fields, ", ")))
+end
+
+print "}"