justinzeus/scholarr
1
0
Fork 0
Code Issues 1 Pull requests 18 Projects Releases Packages Wiki Activity Actions

Initial release prep

Browse source
This commit is contained in:
Justin Visser 2026-02-27 10:46:05 +01:00
parent c9b9d892f4
commit 92395b2b4b
100 changed files with 405 additions and 401 deletions
Download patch file Download diff file Expand all files Collapse all files

6
app/services/scholar/author_rows.py
View file

@ -5,9 +5,9 @@ from html.parser import HTMLParser
from typing import Any
from urllib.parse import parse_qs, urlparse
from app.services.domains.scholar.parser_constants import AUTHOR_SEARCH_MARKER_KEYS
from app.services.domains.scholar.parser_types import ScholarSearchCandidate
from app.services.domains.scholar.parser_utils import (
from app.services.scholar.parser_constants import AUTHOR_SEARCH_MARKER_KEYS
from app.services.scholar.parser_types import ScholarSearchCandidate
from app.services.scholar.parser_utils import (
attr_class,
attr_href,
attr_src,

20
app/services/scholar/parser.py
View file

@ -1,30 +1,30 @@
from __future__ import annotations
from app.services.domains.scholar.author_rows import (
from app.services.scholar.author_rows import (
ScholarAuthorSearchParser,
count_author_search_markers,
)
from app.services.domains.scholar.parser_constants import SCRIPT_STYLE_RE
from app.services.domains.scholar.parser_types import (
from app.services.scholar.parser_constants import SCRIPT_STYLE_RE
from app.services.scholar.parser_types import (
ParsedAuthorSearchPage,
ParsedProfilePage,
PublicationCandidate,
ScholarDomInvariantError,
ScholarMalformedDataError,
)
from app.services.domains.scholar.parser_types import (
from app.services.scholar.parser_types import (
ParseState as ParseState,
)
from app.services.domains.scholar.parser_types import (
from app.services.scholar.parser_types import (
ScholarParserError as ScholarParserError,
)
from app.services.domains.scholar.parser_types import (
from app.services.scholar.parser_types import (
ScholarSearchCandidate as ScholarSearchCandidate,
)
from app.services.domains.scholar.parser_utils import (
from app.services.scholar.parser_utils import (
strip_tags,
)
from app.services.domains.scholar.profile_rows import (
from app.services.scholar.profile_rows import (
count_markers,
extract_articles_range,
extract_profile_image_url,
@ -33,8 +33,8 @@ from app.services.domains.scholar.profile_rows import (
has_show_more_button,
parse_publications,
)
from app.services.domains.scholar.source import FetchResult
from app.services.domains.scholar.state_detection import (
from app.services.scholar.source import FetchResult
from app.services.scholar.state_detection import (
detect_author_search_state,
detect_state,
)

2
app/services/scholar/parser_utils.py
View file

@ -2,7 +2,7 @@ from __future__ import annotations
from html import unescape
from app.services.domains.scholar.parser_constants import TAG_RE
from app.services.scholar.parser_constants import TAG_RE
def normalize_space(value: str) -> str:

6
app/services/scholar/profile_rows.py
View file

@ -5,12 +5,12 @@ from html.parser import HTMLParser
from typing import Any
from urllib.parse import parse_qs, urlparse
from app.services.domains.scholar.parser_constants import (
from app.services.scholar.parser_constants import (
MARKER_KEYS,
SHOW_MORE_BUTTON_RE,
)
from app.services.domains.scholar.parser_types import PublicationCandidate
from app.services.domains.scholar.parser_utils import (
from app.services.scholar.parser_types import PublicationCandidate
from app.services.scholar.parser_utils import (
attr_class,
attr_href,
build_absolute_scholar_url,

2
app/services/scholar/source.py
View file

@ -10,7 +10,7 @@ from urllib.parse import urlencode
from urllib.request import Request, urlopen
from app.logging_utils import structured_log
from app.services.domains.scholar import rate_limit as scholar_rate_limit
from app.services.scholar import rate_limit as scholar_rate_limit
from app.settings import settings
SCHOLAR_PROFILE_URL = "https://scholar.google.com/citations"

6
app/services/scholar/state_detection.py
View file

@ -1,6 +1,6 @@
from __future__ import annotations
from app.services.domains.scholar.parser_constants import (
from app.services.scholar.parser_constants import (
BLOCKED_KEYWORDS,
NETWORK_DNS_ERROR_KEYWORDS,
NETWORK_TIMEOUT_KEYWORDS,
@ -8,8 +8,8 @@ from app.services.domains.scholar.parser_constants import (
NO_AUTHOR_RESULTS_KEYWORDS,
NO_RESULTS_KEYWORDS,
)
from app.services.domains.scholar.parser_types import ParseState, PublicationCandidate, ScholarSearchCandidate
from app.services.domains.scholar.source import FetchResult
from app.services.scholar.parser_types import ParseState, PublicationCandidate, ScholarSearchCandidate
from app.services.scholar.source import FetchResult
def classify_network_error_reason(fetch_error: str | None) -> str: