apache beam docs script

process.py

import fileinput

PATH = "website/www/site/content/en/documentation/programming-guide.md"

seen = []

STATE_DEFAULT = 0
STATE_IN_HIGHLIGHT = 1
STATE_FINISHED_HIGHLIGHT = 2
STATE_AFTER_HIGHLIGHT_NEWLINE = 3

state = STATE_DEFAULT

with fileinput.input(inplace=True, files=(PATH)) as f:
    for line in f:
        if state == STATE_IN_HIGHLIGHT and "{{< /highlight >}}" in line:
          state = STATE_FINISHED_HIGHLIGHT
        elif state == STATE_FINISHED_HIGHLIGHT and line.strip() == "":
          state = STATE_AFTER_HIGHLIGHT_NEWLINE
        elif "{{< highlight" in line:
          if "{{< highlight java >}}" in line:
            seen.append("java")
          elif "{{< highlight py >}}" in line or "{{< highlight python >}}" in line:
            seen.append("py")
          if "{{< highlight go >}}" in line:
            seen.append("go")
          state = STATE_IN_HIGHLIGHT
        elif state == STATE_AFTER_HIGHLIGHT_NEWLINE and line.strip() != "":
          if "py" not in seen:
            print("{{< highlight py />}}\n")
          if "go" not in seen:
            print("{{< highlight go />}}\n")
          seen = []
          state = STATE_DEFAULT
        print(line, end="")
        pass