Removes local connector

This commit is contained in:
Mouse Reeve 2021-09-14 15:26:18 -07:00
parent 357e111411
commit 1f06d1a1d8
10 changed files with 55 additions and 219 deletions

View file

@ -3,4 +3,4 @@ from .settings import CONNECTORS
from .abstract_connector import ConnectorException
from .abstract_connector import get_data, get_image
from .connector_manager import search, local_search, first_search_result
from .connector_manager import search, first_search_result

View file

@ -31,7 +31,6 @@ class AbstractMinimalConnector(ABC):
"isbn_search_url",
"name",
"identifier",
"local",
]
for field in self_fields:
setattr(self, field, getattr(info, field))

View file

@ -55,7 +55,7 @@ def search(query, min_confidence=0.1, return_first=False):
# if we found anything, return it
return result_set[0]
if result_set or connector.local:
if result_set:
results.append(
{
"connector": connector,
@ -71,20 +71,6 @@ def search(query, min_confidence=0.1, return_first=False):
return results
def local_search(query, min_confidence=0.1, raw=False, filters=None):
"""only look at local search results"""
connector = load_connector(models.Connector.objects.get(local=True))
return connector.search(
query, min_confidence=min_confidence, raw=raw, filters=filters
)
def isbn_local_search(query, raw=False):
"""only look at local search results"""
connector = load_connector(models.Connector.objects.get(local=True))
return connector.isbn_search(query, raw=raw)
def first_search_result(query, min_confidence=0.1):
"""search until you find a result that fits"""
return search(query, min_confidence=min_confidence, return_first=True) or None

View file

@ -1,164 +0,0 @@
""" using a bookwyrm instance as a source of book data """
from functools import reduce
import operator
from django.contrib.postgres.search import SearchRank, SearchQuery
from django.db.models import OuterRef, Subquery, F, Q
from bookwyrm import models
from .abstract_connector import AbstractConnector, SearchResult
class Connector(AbstractConnector):
"""instantiate a connector"""
# pylint: disable=arguments-differ
def search(self, query, min_confidence=0, raw=False, filters=None):
"""search your local database"""
filters = filters or []
if not query:
return []
# first, try searching unqiue identifiers
results = search_identifiers(query, *filters)
if not results:
# then try searching title/author
results = search_title_author(query, min_confidence, *filters)
search_results = []
for result in results:
if raw:
search_results.append(result)
else:
search_results.append(self.format_search_result(result))
if len(search_results) >= 10:
break
if not raw:
search_results.sort(key=lambda r: r.confidence, reverse=True)
return search_results
def isbn_search(self, query, raw=False):
"""search your local database"""
if not query:
return []
filters = [{f: query} for f in ["isbn_10", "isbn_13"]]
results = models.Edition.objects.filter(
reduce(operator.or_, (Q(**f) for f in filters))
).distinct()
# when there are multiple editions of the same work, pick the default.
# it would be odd for this to happen.
default_editions = models.Edition.objects.filter(
parent_work=OuterRef("parent_work")
).order_by("-edition_rank")
results = (
results.annotate(
default_id=Subquery(default_editions.values("id")[:1])
).filter(default_id=F("id"))
or results
)
search_results = []
for result in results:
if raw:
search_results.append(result)
else:
search_results.append(self.format_search_result(result))
if len(search_results) >= 10:
break
return search_results
def format_search_result(self, search_result):
cover = None
if search_result.cover:
cover = "%s%s" % (self.covers_url, search_result.cover)
return SearchResult(
title=search_result.title,
key=search_result.remote_id,
author=search_result.author_text,
year=search_result.published_date.year
if search_result.published_date
else None,
connector=self,
cover=cover,
confidence=search_result.rank if hasattr(search_result, "rank") else 1,
)
def format_isbn_search_result(self, search_result):
return self.format_search_result(search_result)
def is_work_data(self, data):
pass
def get_edition_from_work_data(self, data):
pass
def get_work_from_edition_data(self, data):
pass
def get_authors_from_data(self, data):
return None
def parse_isbn_search_data(self, data):
"""it's already in the right format, don't even worry about it"""
return data
def parse_search_data(self, data):
"""it's already in the right format, don't even worry about it"""
return data
def expand_book_data(self, book):
pass
def search_identifiers(query, *filters):
"""tries remote_id, isbn; defined as dedupe fields on the model"""
# pylint: disable=W0212
or_filters = [
{f.name: query}
for f in models.Edition._meta.get_fields()
if hasattr(f, "deduplication_field") and f.deduplication_field
]
results = models.Edition.objects.filter(
*filters, reduce(operator.or_, (Q(**f) for f in or_filters))
).distinct()
if results.count() <= 1:
return results
# when there are multiple editions of the same work, pick the default.
# it would be odd for this to happen.
default_editions = models.Edition.objects.filter(
parent_work=OuterRef("parent_work")
).order_by("-edition_rank")
return (
results.annotate(default_id=Subquery(default_editions.values("id")[:1])).filter(
default_id=F("id")
)
or results
)
def search_title_author(query, min_confidence, *filters):
"""searches for title and author"""
query = SearchQuery(query, config="simple") | SearchQuery(query, config="english")
results = (
models.Edition.objects.filter(*filters, search_vector=query)
.annotate(rank=SearchRank(F("search_vector"), query))
.filter(rank__gt=min_confidence)
.order_by("-rank")
)
# when there are multiple editions of the same work, pick the closest
editions_of_work = results.values("parent_work__id").values_list("parent_work__id")
# filter out multiple editions of the same work
for work_id in set(editions_of_work):
editions = results.filter(parent_work=work_id)
default = editions.order_by("-edition_rank").first()
default_rank = default.rank if default else 0
# if mutliple books have the top rank, pick the default edition
if default_rank == editions.first().rank:
yield default
else:
yield editions.first()

View file

@ -73,19 +73,6 @@ def init_permissions():
def init_connectors():
"""access book data sources"""
Connector.objects.create(
identifier=DOMAIN,
name="Local",
local=True,
connector_file="self_connector",
base_url="https://%s" % DOMAIN,
books_url="https://%s/book" % DOMAIN,
covers_url="https://%s/images/" % DOMAIN,
search_url="https://%s/search?q=" % DOMAIN,
isbn_search_url="https://%s/isbn/" % DOMAIN,
priority=1,
)
Connector.objects.create(
identifier="bookwyrm.social",
name="BookWyrm dot Social",

View file

@ -0,0 +1,17 @@
# Generated by Django 3.2.4 on 2021-09-14 22:10
from django.db import migrations
class Migration(migrations.Migration):
dependencies = [
("bookwyrm", "0096_merge_20210912_0044"),
]
operations = [
migrations.RemoveField(
model_name="connector",
name="local",
),
]

View file

@ -14,7 +14,6 @@ class Connector(BookWyrmModel):
identifier = models.CharField(max_length=255, unique=True)
priority = models.IntegerField(default=2)
name = models.CharField(max_length=255, null=True, blank=True)
local = models.BooleanField(default=False)
connector_file = models.CharField(max_length=255, choices=ConnectorFiles.choices)
api_key = models.CharField(max_length=255, null=True, blank=True)
active = models.BooleanField(default=True)

View file

@ -8,7 +8,24 @@
<ul class="block">
{% for result in local_results.results %}
<li class="pd-4 mb-5">
{% include 'snippets/search_result_text.html' with result=result %}
<div class="columns is-mobile is-gapless">
<div class="column is-cover">
{% include 'snippets/book_cover.html' with book=result cover_class='is-w-xs is-h-xs' %}
</div>
<div class="column is-10 ml-3">
<p>
<strong>
{% include "snippets/book_titleby.html" with book=result %}
</strong>
</p>
<p>
{% if result.first_published_date or result.published_date %}
({% firstof result.first_published_date.year result.published_date.year %})
{% endif %}
</p>
</div>
</div>
</li>
{% endfor %}
</ul>

View file

@ -9,10 +9,8 @@
<strong>
<a
href="{{ result.view_link|default:result.key }}"
{% if remote_result %}
rel="noopener"
target="_blank"
{% endif %}
rel="noopener"
target="_blank"
>{{ result.title }}</a>
</strong>
</p>
@ -26,16 +24,14 @@
{% endif %}
</p>
{% if remote_result %}
<form class="mt-1" action="/resolve-book" method="post">
{% csrf_token %}
<form class="mt-1" action="/resolve-book" method="post">
{% csrf_token %}
<input type="hidden" name="remote_id" value="{{ result.key }}">
<input type="hidden" name="remote_id" value="{{ result.key }}">
<button type="submit" class="button is-small is-link">
{% trans "Import book" %}
</button>
</form>
{% endif %}
<button type="submit" class="button is-small is-link">
{% trans "Import book" %}
</button>
</form>
</div>
</div>

View file

@ -10,6 +10,7 @@ from django.views import View
from bookwyrm import models
from bookwyrm.connectors import connector_manager
from bookwyrm.book_search import search
from bookwyrm.settings import PAGE_LENGTH
from bookwyrm.utils import regex
from .helpers import is_api_request, privacy_filter
@ -31,9 +32,7 @@ class Search(View):
if is_api_request(request):
# only return local book results via json so we don't cascade
book_results = connector_manager.local_search(
query, min_confidence=min_confidence
)
book_results = search(query, min_confidence=min_confidence)
return JsonResponse([r.json() for r in book_results], safe=False)
if query and not search_type:
@ -69,13 +68,13 @@ class Search(View):
def book_search(query, _, min_confidence, search_remote=False):
"""the real business is elsewhere"""
# try a local-only search
if not search_remote:
results = connector_manager.local_search(query, min_confidence=min_confidence)
if results:
# gret, we found something
return [{"results": results}], False
# if there weere no local results, or the request was for remote, search all sources
return connector_manager.search(query, min_confidence=min_confidence), True
results = [{"results": search(query, min_confidence=min_confidence)}]
if results and not search_remote:
return results, False
# if there were no local results, or the request was for remote, search all sources
results += connector_manager.search(query, min_confidence=min_confidence)
return results, True
def user_search(query, viewer, *_):