chore: reformat with black

README.md

@@ -3,7 +3,7 @@
 This is a Python rewrite of Gilles Castel's [Instant Reference](https://github.com/gillescastel/instant-reference) tool.
 (I was not a fan of needing NPM rather than the system package manager to install some dependencies.)
 
-pyinstantref allows you to copy a link to a specific page in a PDF with a single keybind in Zathura.
+pyinstantref allows you to copy a link to a specific page or header in a PDF with a single keybind in Zathura.
 You can then paste this reference in your notes and other documents.
 
 For now, it only works with my own [templates](https://github.com/dogeystamp/typst-templates) for [Typst](https://github.com/typst/typst),
@@ -50,7 +50,7 @@ This will make Ctrl-L copy a reference to the current page in Zathura.
 
 ## limitations
 
-Currently, the following features are missing:
+Currently, the following features are missing compared to Castel's version:
 - ArXiv support
 - LaTeX output
 - Support for other PDF readers (e.g. Evince)
@@ -58,3 +58,9 @@ Currently, the following features are missing:
 Feel free to send pull requests,
 although this project is primarily for my own usage
 and I can not make any guarantees.
+
+Also:
+- Section references are unreliable because titles might change,
+  and there might be sections with the same title.
+  Proper IDs for bookmarks are possible,
+  but not until Typst resolves [issue #1352](https://github.com/typst/typst/issues/1352).

copy_ref

@@ -1,10 +1,10 @@
 #!/usr/bin/env python3
 
-from pdf_data import get_page_pdf
+from pdf_data import get_page_pdf, get_section_pdf
 from datatypes import *
 from enum import Enum, auto
+from util import notify
 import subprocess
-import pydbus
 import argparse
 import formatter.typst as typst_fmt
 
@@ -32,21 +32,19 @@ def copy_ref(ref: Reference, format: LinkFormat) -> None:
     clip_copy(link_txt)
 
 
-def notify(title:str, txt: str) -> None:
-    """Send a text notification."""
-    bus = pydbus.SessionBus()
-    notifs = bus.get(".Notifications")
-    notifs.Notify("instantref", 0, "dialog-information", title, txt, [], {}, 5000)
-
-
 if __name__ == "__main__":
     args = parser.parse_args()
 
     if args.section:
-        raise NotImplementedError("--section isn't implemented")
-
-    ref = get_page_pdf()
+        ref = get_section_pdf()
+    else:
+        ref = get_page_pdf()
 
     format = LinkFormat.TYPST
     copy_ref(ref, format)
-    notify("Copied ref", f"{ref.filepath.name} p. {ref.page}")
+
+    match ref:
+        case PDFPage():
+            notify("Copied ref", f"{ref.filepath.name} p. {ref.page}")
+        case PDFSection():
+            notify("Copied ref", f"{ref.filepath.name} sec. {ref.title}")

datatypes.py

@@ -11,6 +11,7 @@ ProcessId = NewType("ProcessId", int)
 @dataclass
 class _Reference:
     """Reference to a location within a file."""
+
     pass
 
 
@@ -63,3 +64,11 @@ PDFReference = Union[PDFPage, PDFSection]
 # for now no other format is implemented
 # replace this with an union if that happens
 Reference = PDFReference
+
+
+# PyMuPDF type
+@dataclass
+class FitzBookmark:
+    level: int
+    title: SectionTitle
+    page: PageNumber

formatter/typst.py

@@ -4,6 +4,7 @@ from datatypes import PDFPage, PDFSection, PDFReference, Reference
 from typing import assert_never
 from pathlib import Path
 
+
 def format_pdf_link(ref: PDFReference) -> str:
     path_str = environ.get("TYPST_ROOT", None)
     if path_str is None:
@@ -19,19 +20,24 @@ def format_pdf_link(ref: PDFReference) -> str:
         format_path = str(ref.filepath.absolute())
 
     params = {}
+    default_label = ""
 
     match ref:
         case PDFPage():
             params["page"] = ref.page
         case PDFSection():
             params["section"] = ref.title
+            default_label = ref.title
         case _ as obj:
             assert_never(obj)
 
     if relative:
-        return f'#lref("{format_path}?{urlencode(params)}", pdfref: true)[]'
+        return (
+            f'#lref("{format_path}?{urlencode(params)}", pdfref: true)[{default_label}]'
+        )
     else:
-        return f'#link("pdfref://{format_path}?{urlencode(params)}")[]'
+        return f'#link("pdfref://{format_path}?{urlencode(params)}")[{default_label}]'
+
 
 def ref(ref: Reference) -> str:
     """Formats a Reference."""

pdf_data.py

@@ -1,7 +1,24 @@
 from pathlib import Path
 from datatypes import *
+from typing import cast, Any
+from util import rofi
 import pydbus
 import subprocess
+import fitz
+
+
+def get_section_pdf() -> PDFSection:
+    page_ref: PDFPage = get_page_pdf()
+    with fitz.Document(page_ref.filepath) as doc:
+        toc = [FitzBookmark(*x) for x in cast(Any, doc).get_toc()]
+        page_headers = [x for x in toc if x.page == page_ref.page]
+
+        rofi_res = rofi([f"{x.title}" for x in page_headers], prompt="Select header: ")
+        if rofi_res is None or rofi_res.index is None:
+            raise RuntimeError("No header was selected.")
+        selected_header = page_headers[rofi_res.index]
+
+        return PDFSection(filepath=page_ref.filepath, title=selected_header.title)
 
 
 def get_page_pdf() -> PDFPage:

pdfref_handler

@@ -1,15 +1,31 @@
 #!/usr/bin/env python3
 # pdfref:// URL handler
 
-import subprocess
 from urllib.parse import urlparse, parse_qs
 from sys import argv
-
-from datatypes import PageNumber
+from datatypes import *
+from typing import cast, Any
+from util import notify
+import subprocess
+import fitz
 
 url = urlparse(argv[1])
 query = parse_qs(url.query)
 
 page: PageNumber = PageNumber(int(query.get("page", ["0"])[0]))
+section: SectionTitle = SectionTitle(query.get("section", [])[0])
+
+if section != []:
+    with fitz.Document(url.path) as doc:
+        toc = [FitzBookmark(*x) for x in cast(Any, doc).get_toc()]
+        headers = [x for x in toc if x.title == section]
+        if headers == []:
+            notify("", f"Failed to find section '{section}': did the title change?")
+        else:
+            if len(headers) > 1:
+                notify("", f"Multiple sections '{section}' found: page might be incorrect")
+
+            page = headers[0].page
+
 
 subprocess.run(["zathura", "--page", str(page), url.path], text=True)

requirements.txt

@@ -1,3 +1,4 @@
 pycairo==1.24.0
 pydbus==0.6.0
 PyGObject==3.44.1
+PyMuPDF==1.22.5

util.py

@@ -1,7 +1,9 @@
 import subprocess
+import pydbus
 from dataclasses import dataclass
 from typing import Optional
 
+
 @dataclass
 class RofiResult:
     """Data returned from Rofi.
@@ -17,12 +19,15 @@ class RofiResult:
     custom_bind
         ID of custom bind used to select entry. None if no custom bind was used.
     """
+
     index: Optional[int]
     value: str
     custom_bind: Optional[int]
 
 
-def rofi(entries: list[str], prompt: str="> ", fuzzy=True, extra_args=[]) -> Optional[RofiResult]:
+def rofi(
+    entries: list[str], prompt: str = "> ", fuzzy=True, extra_args=[]
+) -> Optional[RofiResult]:
     """Start a Rofi prompt.
 
     Returns
@@ -38,8 +43,10 @@ def rofi(entries: list[str], prompt: str="> ", fuzzy=True, extra_args=[]) -> Opt
     args += extra_args
 
     ret = RofiResult(None, "", None)
-    
-    res = subprocess.run(args, input="\0".join(entries), stdout=subprocess.PIPE, text=True)
+
+    res = subprocess.run(
+        args, input="\0".join(entries), stdout=subprocess.PIPE, text=True
+    )
     match res.returncode:
         case 0:
             pass
@@ -56,3 +63,10 @@ def rofi(entries: list[str], prompt: str="> ", fuzzy=True, extra_args=[]) -> Opt
         pass
 
     return ret
+
+
+def notify(title: str, txt: str) -> None:
+    """Send a text notification."""
+    bus = pydbus.SessionBus()
+    notifs = bus.get(".Notifications")
+    notifs.Notify("instantref", 0, "dialog-information", title, txt, [], {}, 5000)