qutebrowser: Add doi2scihub script

Added script which takes you to the corresponding sci-hub entry for any
DOI. DOIs can be passed in three ways:

- via hinted link (shortcut `;p` to start hinting)
- via selected text (select text then invoke `send-to-scihub` command
  with `"p`)
- or from meta tags in current page (invoke `send-to-scihub` command
  with `"p` when on article page)

It will grab the newest sci-hub link and attempt to bring you the
corresponding pdf file.

qutebrowser/.config/qutebrowser/alias.py

@@ -16,6 +16,8 @@ c.aliases["send-to-archive"] = "open https://web.archive.org/web/{url}"
 
 # save current page to pdf file
 c.aliases["save_to_pdf"] = "spawn --userscript pagetopdf.sh"
+# open sci-hub pdf for doi
+c.aliases["send-to-scihub"] = "spawn --userscript doi2scihub"
 
 # translate current page / selection with google translate
 c.aliases["translate-page-google"] = "spawn --userscript translate_google.sh"

qutebrowser/.config/qutebrowser/maps.py

@@ -77,6 +77,8 @@ config.bind('"T', "translate-selection-google", mode="normal")
 config.bind('"q', "show-qr")
 
 config.bind(lleader + "r", "spawn --userscript readability")
+config.bind('"p', "send-to-scihub", mode="normal")  # view current page doi on scihub
+config.bind(";p", "hint links run send-to-scihub")  # view linked doi on scihub
 
 # set stylesheets for the browser to use
 config.bind(

qutebrowser/.config/qutebrowser/url.py

@@ -1,7 +1,7 @@
 from qutebrowser.api import interceptor
 
 c.url.searchengines = {
-    "#sci": "https://sci-hub.do/{}",
+    "sci": "https://sci-hub.ru/{}",
     "DEFAULT": "https://search.martyoeh.me/?q={}",
     "al": "https://wiki.archlinux.org/index.php/{}",
     "alt": "https://alternativeto.net/software/{}/?license=opensource",

qutebrowser/.local/share/qutebrowser/userscripts/doi2scihub

@@ -0,0 +1,105 @@
+#!/usr/bin/env python3
+"""
+Goes to the sci-hub page for the current article, based on DOI.
+
+Can be invoked with DOI on a page selected,
+through the hinting mode when selecting a DOI link
+or on a publisher page (any page where doi meta-tags are set) -
+works on ScienceDirect, Taylor&Francis, Springer, etc.
+
+Updates its sci-hub link based on the one listed on sci-hub wiki page.
+
+Based on the work in
+https://github.com/cadadr/configuration/blob/4b6a241d04d113f322b960890a0d0a0ab783a7b3/dotfiles/qutebrowser/userscripts/doi
+with much gratitude.
+"""
+
+
+import os
+import re
+import sys
+import html.parser
+import requests
+
+
+mode = os.getenv("QUTE_MODE")
+
+text = None
+
+
+class DoiTagParser(html.parser.HTMLParser):
+    doi = None
+
+    def handle_starttag(self, tag: str, attrs: list[tuple[str, str | None]]) -> None:
+        if self.doi == None and tag == "meta":
+            if (
+                ("name", "citation_doi") in attrs
+                or ("name", "dc.identifier") in attrs
+                or ("scheme", "doi") in attrs
+            ):
+                for att in attrs:
+                    if att[0] == "content":
+                        self.doi = att[1]
+                        break
+
+
+class SciHubLinkParser(html.parser.HTMLParser):
+    current = None
+    link_patt = re.compile(r"^(?P<url>https?://sci-hub\..+)/about$")
+
+    def handle_starttag(self, tag: str, attrs: list[tuple[str, str | None]]) -> None:
+        if self.current == None and tag == "a":
+            for att in attrs:
+                if att[0] == "href" and self.link_patt.match(att[1] or ""):
+                    match = self.link_patt.match(att[1] or "")
+                    self.current = match["url"] if match and match["url"] else None
+
+
+def get_scihub_url(wiki_page: str = "https://wikiless.org/wiki/Sci-Hub"):
+    resp = requests.get(wiki_page)
+    parser = SciHubLinkParser()
+    parser.feed(resp.text)
+    return parser.current or "https://sci-hub.ru"
+
+
+if mode == "hints":
+    text = os.getenv("QUTE_URL", "").strip()
+elif mode == "command" and os.getenv("QUTE_SELECTED_TEXT"):
+    text = os.getenv("QUTE_SELECTED_TEXT", "").strip()
+elif os.getenv("QUTE_HTML"):
+    # TODO implement html source-based doi search for current page here
+    # use python htmlparser and find metatags: e.g. citation_doi, dc.identifier
+    with open(os.getenv("QUTE_HTML", ""), "r") as source:
+        parser = DoiTagParser()
+        parser.feed(source.read())
+        text = parser.doi
+
+with open(os.getenv("QUTE_FIFO", ""), "w") as fifo:
+    if not text:
+        fifo.write(f'message-warning "Could not find a valid DOI"')
+        sys.exit()
+
+    # DOI syntax: https://www.doi.org/doi_handbook/2_Numbering.html#2.2.
+    #
+    # Note that this probably matches a subset of possible DOIs, as it
+    # seems that there’s no practical limitation on neither the length nor
+    # the contents of the DOI.  But IMHO this is a healthy subset.
+    doi_re = re.compile(
+        # match possible URI prefix
+        r"(?P<blah>((https?)?://)?doi\.org/)?"
+        # match actual DOI
+        r"(?P<meat>[a-zA-Z0-9\./\-_]+)"
+    )
+
+    match = doi_re.match(text)
+
+    if match is None or match["meat"] is None:
+        fifo.write(
+            f"message-warning \"'{text}' is probably not a DOI, or update regexp\""
+        )
+
+    else:
+        url = get_scihub_url()
+        doi = match["meat"]
+
+        fifo.write(f"open -t {url}/{doi}")