Update ReadMe.md

ReadMe.md

@@ -1,3 +1,10 @@
+# yt-transcript-tools
+Getting transcripts into epub format.
+
+Download a transcript, place it in this directory, run
+`./convert_transcript.lua FILE`, fill in author and link as
+requested, and you will have a usable epub.
+
 This is heavily WIP.
 
 I used the apply_margin script on a transcript downloaded from
@@ -14,3 +21,9 @@ the export will be different enough to fuck up the margin fix.
 I am really frustrated because pandoc doesn't support plain text.
 I went through many tests and iterations before coming to even
 this minimal start and it was extremely frustrating.
+
+## Tasks
+- [x] I need an option / 2nd script to output a version with no
+  timecode formatting for easier text extraction.
+  - I made it actually just combine both in the same document,
+    easier to manage that way.

apply_margin.lua

@@ -1,16 +0,0 @@
-#!/usr/bin/env luajit
-local a, b = arg[1], arg[2]
-
-local file = io.open(b, "w")
-assert(file, "Couldn't open \"" .. b .. "\"")
-
-for line in io.lines(a) do
-  if not (line:sub(3, 3) == ":") then
-    file:write(string.rep(" ", 6))
-  end
-
-  file:write(line)
-  file:write("\n")
-end
-
-file:close()

convert_transcript.lua

@@ -1,5 +1,5 @@
 #!/usr/bin/env luajit
-local version = "0.1"
+local version = "0.4.0"
 
 local input_file_name = arg[1]
 assert(input_file_name,
@@ -26,29 +26,43 @@ if link and #link > 0 then
   frontmatter[#frontmatter + 1] = "From [" .. link .. "](" .. link .. ")\n"
 end
 
+local input_lines = {}
+for line in io.lines(input_file_name) do
+  input_lines[#input_lines + 1] = line
+end
+
+local output_lines = {}
+local timecode_free_lines = {}
+for index, line in ipairs(input_lines) do
+  if not (line:sub(3, 3) == ":") then
+    local previous_line = input_lines[index - 1]
+    if previous_line and previous_line:sub(3, 3) == ":" then
+      output_lines[#output_lines] = "```\n" .. previous_line
+      output_lines[#output_lines + 1] = string.rep(" ", 6) .. line .. "\n```"
+    else
+      output_lines[#output_lines + 1] = "`" .. string.rep(" ", 6) .. line .. "`"   -- pandoc will ignore this margin
+    end
+    timecode_free_lines[#timecode_free_lines + 1] = line
+  else
+    output_lines[#output_lines + 1] = "`" .. line .. "`"
+    timecode_free_lines[#timecode_free_lines + 1] = line:sub(7)
+  end
+end
+
 local output_file = io.open(output_file_name, "w")
 assert(output_file, "Could not open \"" .. output_file_name .. "\"")
 
-for _, line in ipairs(frontmatter) do
-  output_file:write(line)
-  output_file:write("\n")
-end
+output_file:write(table.concat(frontmatter, "\n"))
+output_file:write("\n")
 
-for line in io.lines(input_file_name) do
-  output_file:write("`") -- make this line a code block on its own
-
-  -- if not appearing to start with a timecode,
-  -- indent by the width of a timecode
-  if not (line:sub(3, 3) == ":") then
-    output_file:write(string.rep(" ", 6))
-  end
-
-  output_file:write(line)
-  output_file:write("`") -- close the code block
-  output_file:write("\n")
-end
+output_file:write("# Raw Transcript\n\n")
+output_file:write(table.concat(timecode_free_lines, "\n"))
+output_file:write("\n\n")
 
+output_file:write("# Timecoded Transcript\n\n")
+output_file:write(table.concat(output_lines, "\n"))
+output_file:write("\n")
 output_file:close()
 
 os.execute("pandoc --from markdown+hard_line_breaks \""
-  .. output_file_name .. "\" -o \"" .. title .. ".epub\"")
+  .. output_file_name .. "\" --toc=true -o \"" .. title .. ".epub\"")

oops_all_code.lua

@@ -1,14 +0,0 @@
-#!/usr/bin/env luajit
-local a, b = arg[1], arg[2]
-
-local file = io.open(b, "w")
-assert(file, "Couldn't open \"" .. b .. "\"")
-
-for line in io.lines(a) do
-  file:write("`")
-  file:write(line)
-  file:write("`")
-  file:write("\n")
-end
-
-file:close()