JetCore/jetc/strip_fragments.py

"""
Strips the dead SPA-fragment dual-render wrapper from a barant-site/jet/*.jet
route file, leaving just the real page content (the former "_full_NAME" tag
body), unwrapped, with no fragment-flag try/catch, no _frag_NAME tag
definition, no dispatch <if>, and no dead client-side SPA navigation script.

Structure being removed (present in all 20 files, verified consistent):

  #api GET /NAME
  <try name="_fragerr">...<catch>...</try><tag name="_frag_NAME">
      ... fragment-only content, NEVER actually invoked any more since the
      client-side SPA click handlers were already disabled earlier this
      session ...
  </tag><tag name="_full_NAME">
      ... the REAL page content - full doctype/head/style/body wrapper ...
      <script>
      (function() {
         var FRAGMENT_EXCLUDED = [...];
         function pageNameFromPath(...) {...}
         function loadPage(...) {...}
         // SPA click/popstate interception disabled for now ...
      })();
      </script></body></html>
  </tag><if value1="$[fragflag]" value2="1" type="eq"><_frag_NAME /><else><_full_NAME /></if>

Kept: "#api GET /NAME\n" plus everything between "<tag name=\"_full_NAME\">"
and its own matching close, MINUS the dead <script> block.
"""
import re
import sys


def strip_dead_spa_script(content):
    pattern = re.compile(
        r"<script>\s*\(function\(\)\s*\{.*?SPA click/popstate interception disabled.*?\}\)\(\);\s*</script>",
        re.DOTALL,
    )
    new_content, n = pattern.subn("", content, count=1)
    return new_content, n


def strip_fragment_wrapper(content, page_name):
    lines = content.split("\n", 1)
    assert lines[0].startswith("#api GET /"), f"unexpected first line: {lines[0]!r}"
    api_line = lines[0]
    rest = lines[1] if len(lines) > 1 else ""

    start_marker = (
        '<try name="_fragerr"><set name="fragflag" expr="$[fragment]" />'
        '<catch><set name="fragflag" value="" /></try>'
        f'<tag name="_frag_{page_name}">'
    )
    if start_marker not in rest:
        return None, f"start marker not found"
    idx = rest.index(start_marker)
    prefix_extra = rest[:idx]  # should just be leading whitespace/newline, kept as-is
    rest = rest[idx + len(start_marker):]

    mid_marker = f'</tag><tag name="_full_{page_name}">'
    if mid_marker not in rest:
        return None, f"mid marker not found"
    mid_idx = rest.index(mid_marker)
    rest = rest[mid_idx + len(mid_marker):]

    end_pattern = re.compile(
        r'</tag><if value1="\$\[fragflag\]" value2="1" type="eq">'
        rf'<_frag_{re.escape(page_name)} /><else><_full_{re.escape(page_name)} /></if>\s*$'
    )
    m = end_pattern.search(rest)
    if not m:
        return None, f"end marker not found"
    full_content = rest[:m.start()]

    full_content, nscript = strip_dead_spa_script(full_content)

    result = api_line + "\n" + prefix_extra + full_content
    return result, f"ok (removed {nscript} dead script block(s))"


def main():
    if len(sys.argv) != 4:
        print("usage: strip_fragments.py <input> <output> <page_name>")
        sys.exit(1)
    with open(sys.argv[1], encoding="utf-8", newline="") as f:
        content = f.read()
    result, status = strip_fragment_wrapper(content, sys.argv[3])
    if result is None:
        print(f"FAILED for {sys.argv[3]}: {status}")
        sys.exit(1)
    with open(sys.argv[2], "w", encoding="utf-8", newline="") as f:
        f.write(result)
    print(f"{sys.argv[3]}: {status} ({len(content)} -> {len(result)} bytes)")


if __name__ == "__main__":
    main()