implement sample lua filter

jimmy_cli.spec

@@ -1,9 +1,10 @@
 # -*- mode: python ; coding: utf-8 -*-
 from PyInstaller.utils.hooks import collect_data_files
 
-# pypandoc: https://github.com/orgs/pyinstaller/discussions/8387
-datas = [(".version", ".")]
+# https://pyinstaller.org/en/stable/spec-files.html#adding-data-files
+datas = [(".version", "."), ("src/pandoc_filter/*.lua", "src/pandoc_filter")]
 datas += collect_data_files("anyblock_exporter")
+# pypandoc: https://github.com/orgs/pyinstaller/discussions/8387
 datas += collect_data_files("pypandoc")
 
 

src/common.py

@@ -8,6 +8,7 @@
 from pathlib import Path
 import pkgutil
 import random
+import sys
 import tarfile
 import tempfile
 import time
@@ -52,6 +53,12 @@ def wrapper(*args, **kwargs):
     return cast(F, wrapper)
 
 
+# Pyinstaller has different root path than module.
+# https://stackoverflow.com/a/44352931/7410886
+ROOT_PATH = Path(getattr(sys, "_MEIPASS", Path(__file__).parent.parent))
+PANDOC_FILTER_PATH = ROOT_PATH / "src/pandoc_filter"
+
+
 def safe_path(path: Path | str, max_name_length: int = 50) -> Path | str:
     r"""
     Return a safe version of the provided path or string.

src/formats/synology_note_station.py

@@ -71,12 +71,6 @@ def streamline_html(content_html: str) -> str:
             iframe.string = iframe.attrs["src"]
         iframe.attrs = {"href": iframe.attrs["src"]}
 
-    # hack: In the original data, the attachment_id is stored in the
-    # "ref" attribute. Mitigate by storing it in the "src" attribute.
-    for img in soup.find_all("img"):
-        if (new_src := img.attrs.get("ref")) is not None:
-            img.attrs["src"] = new_src
-
     return str(soup)
 
 
@@ -203,7 +197,9 @@ def convert_note(self, note_id, note_id_title_map):
         note_links: imf.NoteLinks = []
         if (content_html := note.get("content")) is not None:
             content_html = streamline_html(content_html)
-            content_markdown = markdown_lib.common.markup_to_markdown(content_html)
+            content_markdown = markdown_lib.common.markup_to_markdown(
+                content_html, filters=["synology_note_station.lua"]
+            )
             content_markdown = content_markdown.replace("{TEMPORARYNEWLINE}", "<br>")
             # note title only needed for debug message
             body, resources_referenced, note_links = self.handle_markdown_links(

src/jimmy.py

@@ -2,8 +2,6 @@
 
 import importlib
 import logging
-from pathlib import Path
-import sys
 
 import pypandoc
 from rich import print  # pylint: disable=redefined-builtin
@@ -97,10 +95,7 @@ def get_tree(root_notebooks: imf.Notebooks, root_tree: Tree) -> Tree:
 
 
 def get_jimmy_version():
-    # Pyinstaller has different path than module.
-    # https://stackoverflow.com/a/44352931/7410886
-    base_path = getattr(sys, "_MEIPASS", Path(__file__).parent)
-    version_file = Path(base_path) / ".version"
+    version_file = common.ROOT_PATH / ".version"
     return (
         version_file.read_text().lstrip("v").rstrip()
         if version_file.is_file()

src/markdown_lib/common.py

@@ -10,6 +10,7 @@
 from markdown.extensions import Extension
 import pypandoc
 
+import common
 
 LOGGER = logging.getLogger("jimmy")
 
@@ -218,20 +219,29 @@ def get_inline_tags(text: str, start_characters: list[str]) -> list[str]:
 # fmt:on
 
 
-def markup_to_markdown(text: str, format_: str = "html") -> str:
+def markup_to_markdown(
+    text: str, format_: str = "html", filters: list[str] | None = None
+) -> str:
+    if filters is None:
+        filters = []
     text_md = pypandoc.convert_text(
         text,
         PANDOC_OUTPUT_FORMAT,
         format=format_,
         sandbox=True,
         extra_args=["--wrap=none"],
+        filters=[str(common.PANDOC_FILTER_PATH / f) for f in filters],
     )
     if "[TABLE]" in text_md:
         LOGGER.warning("Table is too complex and can't be converted to markdown.")
     return text_md.strip()
 
 
-def file_to_markdown(file_: Path, resource_folder: Path) -> str:
+def file_to_markdown(
+    file_: Path, resource_folder: Path, filters: list[str] | None = None
+) -> str:
+    if filters is None:
+        filters = []
     file_md = pypandoc.convert_file(
         file_,
         PANDOC_OUTPUT_FORMAT,
@@ -242,6 +252,7 @@ def file_to_markdown(file_: Path, resource_folder: Path) -> str:
             # don't create artificial line breaks
             "--wrap=none",
         ],
+        filters=[str(common.PANDOC_FILTER_PATH / f) for f in filters],
     )
     if "[TABLE]" in file_md:
         LOGGER.warning("Table is too complex and can't be converted to markdown.")

src/pandoc_filter/synology_note_station.lua

@@ -0,0 +1,7 @@
+function Image(img)
+    -- In the original data, the "src" is stored in the "ref" attribute. Move it where it belongs.
+    if img.attributes["ref"] then
+        img.src = img.attributes["ref"]
+    end
+    return img
+end