Added feature to treat couples of languages as equal when searching for subtitles

* Add 'Language-equals' support

This feature will treat couples of languages as equal for list-subtitles
operations. It's optional; its methods won't do anything if an empy list
is set. See more info at docstrings from 'subliminal_patch.core'.

For example, let's say I only want to have "Spanish (es.srt)" subtitles
and I don't care about the differences between Spain and LATAM spanish.
This feature will allow me to always get European Spanish even from
LATAM Spanish providers like Argenteam and Subdivx.

Example for config.ini:

language_equals = ['spa-MX:spa']

(Which means all Latam Spanish subtitles from every provider will be
converted to European Spanish)

* Add PT and ZH language tests

* Add HI and Forced parsing for language pairs

Format example: ["en@HI:en", "es-MX@forced:es-MX"]

* Update languages.py

* Update API definition to reflect the previous change

* Add language equals table to the UI (test only)

* Add global language selector and get language from code3 utilities

* Add unit tests for language equal feature

* Add encode function to language equal feature

* Add CRUD methods to the language equals panel

* Add equals description

* Add parsing support for alpha3 custom languages

* no log: add more tests

* Add forced and hi support to the language equal target

---------

Co-authored-by: morpheus65535 <louis_vezina@hotmail.com>
Co-authored-by: LASER-Yi <liangyi0007@gmail.com>

bazarr/api/system/languages.py

@@ -4,7 +4,7 @@ from flask_restx import Resource, Namespace, reqparse
 from operator import itemgetter
 
 from app.database import TableHistory, TableHistoryMovie, TableSettingsLanguages
-from languages.get_languages import alpha2_from_alpha3, language_from_alpha2
+from languages.get_languages import alpha2_from_alpha3, language_from_alpha2, alpha3_from_alpha2
 
 from ..utils import authenticate, False_Keys
 
@@ -46,6 +46,7 @@ class Languages(Resource):
                     try:
                         languages_dicts.append({
                             'code2': code2,
+                            'code3': alpha3_from_alpha2(code2),
                             'name': language_from_alpha2(code2),
                             # Compatibility: Use false temporarily
                             'enabled': False
@@ -55,6 +56,7 @@ class Languages(Resource):
         else:
             languages_dicts = TableSettingsLanguages.select(TableSettingsLanguages.name,
                                                             TableSettingsLanguages.code2,
+                                                            TableSettingsLanguages.code3,
                                                             TableSettingsLanguages.enabled)\
                 .order_by(TableSettingsLanguages.name).dicts()
             languages_dicts = list(languages_dicts)

bazarr/app/config.py

@@ -83,7 +83,8 @@ defaults = {
         'default_und_audio_lang': '',
         'default_und_embedded_subtitles_lang': '',
         'parse_embedded_audio_track': 'False',
-        'skip_hashing': 'False'
+        'skip_hashing': 'False',
+        'language_equals': '[]',
     },
     'auth': {
         'type': 'None',
@@ -300,7 +301,8 @@ array_keys = ['excluded_tags',
               'excluded_series_types',
               'enabled_providers',
               'path_mappings',
-              'path_mappings_movie']
+              'path_mappings_movie',
+              'language_equals']
 
 str_keys = ['chmod']
 

bazarr/app/get_providers.py

@@ -20,6 +20,7 @@ from subliminal_patch.extensions import provider_registry
 
 from app.get_args import args
 from app.config import settings, get_array_from
+from languages.get_languages import CustomLanguage
 from app.event_handler import event_stream
 from utilities.binaries import get_binary
 from radarr.blacklist import blacklist_log_movie
@@ -115,6 +116,49 @@ def provider_pool():
     return subliminal_patch.core.SZProviderPool
 
 
+def _lang_from_str(content: str):
+    " Formats: es-MX en@hi es-MX@forced "
+    extra_info = content.split("@")
+    if len(extra_info) > 1:
+        kwargs = {extra_info[-1]: True}
+    else:
+        kwargs = {}
+
+    content = extra_info[0]
+
+    try:
+        code, country = content.split("-")
+    except ValueError:
+        lang = CustomLanguage.from_value(content)
+        if lang is not None:
+            lang = lang.subzero_language()
+            return lang.rebuild(lang, **kwargs)
+
+        code, country = content, None
+
+    return subliminal_patch.core.Language(code, country, **kwargs)
+
+
+def get_language_equals(settings_=None):
+    settings_ = settings_ or settings
+
+    equals = get_array_from(settings_.general.language_equals)
+    if not equals:
+        return []
+
+    items = []
+    for equal in equals:
+        try:
+            from_, to_ = equal.split(":")
+            from_, to_ = _lang_from_str(from_), _lang_from_str(to_)
+        except Exception as error:
+            logging.info("Invalid equal value: '%s' [%s]", equal, error)
+        else:
+            items.append((from_, to_))
+
+    return items
+
+
 def get_providers():
     providers_list = []
     existing_providers = provider_registry.names()

bazarr/subtitles/pool.py

@@ -8,7 +8,7 @@ from inspect import getfullargspec
 
 from radarr.blacklist import get_blacklist_movie
 from sonarr.blacklist import get_blacklist
-from app.get_providers import get_providers, get_providers_auth, provider_throttle, provider_pool
+from app.get_providers import get_providers, get_providers_auth, provider_throttle, provider_pool, get_language_equals
 
 from .utils import get_ban_list
 
@@ -19,10 +19,11 @@ def _init_pool(media_type, profile_id=None, providers=None):
     return pool(
         providers=providers or get_providers(),
         provider_configs=get_providers_auth(),
-        blacklist=get_blacklist() if media_type == 'series' else get_blacklist_movie(),
+        blacklist=get_blacklist() if media_type == "series" else get_blacklist_movie(),
         throttle_callback=provider_throttle,
         ban_list=get_ban_list(profile_id),
         language_hook=None,
+        language_equals=get_language_equals(),
     )
 
 
@@ -54,8 +55,19 @@ def _update_pool(media_type, profile_id=None):
     return pool.update(
         get_providers(),
         get_providers_auth(),
-        get_blacklist() if media_type == 'series' else get_blacklist_movie(),
+        get_blacklist() if media_type == "series" else get_blacklist_movie(),
         get_ban_list(profile_id),
+        get_language_equals(),
+    )
+
+
+def _pool_update(pool, media_type, profile_id=None):
+    return pool.update(
+        get_providers(),
+        get_providers_auth(),
+        get_blacklist() if media_type == "series" else get_blacklist_movie(),
+        get_ban_list(profile_id),
+        get_language_equals(),
     )
 
 

frontend/src/components/bazarr/LanguageSelector.tsx

@@ -0,0 +1,34 @@
+import { useLanguages } from "@/apis/hooks";
+import { Selector, SelectorProps } from "@/components/inputs";
+import { useSelectorOptions } from "@/utilities";
+import { FunctionComponent, useMemo } from "react";
+
+interface LanguageSelectorProps
+  extends Omit<SelectorProps<Language.Server>, "options" | "getkey"> {
+  enabled?: boolean;
+}
+
+const LanguageSelector: FunctionComponent<LanguageSelectorProps> = ({
+  enabled = false,
+  ...selector
+}) => {
+  const { data } = useLanguages();
+
+  const filteredData = useMemo(() => {
+    if (enabled) {
+      return data?.filter((value) => value.enabled);
+    } else {
+      return data;
+    }
+  }, [data, enabled]);
+
+  const options = useSelectorOptions(
+    filteredData ?? [],
+    (value) => value.name,
+    (value) => value.code3
+  );
+
+  return <Selector {...options} searchable {...selector}></Selector>;
+};
+
+export default LanguageSelector;

frontend/src/pages/Settings/Languages/equals.test.ts

@@ -0,0 +1,196 @@
+import {
+  decodeEqualData,
+  encodeEqualData,
+  LanguageEqualData,
+  LanguageEqualImmediateData,
+} from "@/pages/Settings/Languages/equals";
+import { describe, expect, it } from "vitest";
+
+describe("Equals Parser", () => {
+  it("should parse from string correctly", () => {
+    interface TestData {
+      text: string;
+      expected: LanguageEqualImmediateData;
+    }
+
+    function testParsedResult(
+      text: string,
+      expected: LanguageEqualImmediateData
+    ) {
+      const result = decodeEqualData(text);
+
+      if (result === undefined) {
+        expect(false, `Cannot parse '${text}' as language equal data`);
+        return;
+      }
+
+      expect(
+        result,
+        `${text} does not match with the expected equal data`
+      ).toStrictEqual(expected);
+    }
+
+    const testValues: TestData[] = [
+      {
+        text: "spa-MX:spa",
+        expected: {
+          source: {
+            content: "spa-MX",
+            hi: false,
+            forced: false,
+          },
+          target: {
+            content: "spa",
+            hi: false,
+            forced: false,
+          },
+        },
+      },
+      {
+        text: "zho@hi:zht",
+        expected: {
+          source: {
+            content: "zho",
+            hi: true,
+            forced: false,
+          },
+          target: {
+            content: "zht",
+            hi: false,
+            forced: false,
+          },
+        },
+      },
+      {
+        text: "es-MX@forced:es-MX",
+        expected: {
+          source: {
+            content: "es-MX",
+            hi: false,
+            forced: true,
+          },
+          target: {
+            content: "es-MX",
+            hi: false,
+            forced: false,
+          },
+        },
+      },
+      {
+        text: "en:en@hi",
+        expected: {
+          source: {
+            content: "en",
+            hi: false,
+            forced: false,
+          },
+          target: {
+            content: "en",
+            hi: true,
+            forced: false,
+          },
+        },
+      },
+    ];
+
+    testValues.forEach((data) => {
+      testParsedResult(data.text, data.expected);
+    });
+  });
+
+  it("should encode to string correctly", () => {
+    interface TestData {
+      source: LanguageEqualData;
+      expected: string;
+    }
+
+    const testValues: TestData[] = [
+      {
+        source: {
+          source: {
+            content: {
+              name: "Abkhazian",
+              code2: "ab",
+              code3: "abk",
+              enabled: false,
+            },
+            hi: false,
+            forced: false,
+          },
+          target: {
+            content: {
+              name: "Aragonese",
+              code2: "an",
+              code3: "arg",
+              enabled: false,
+            },
+            hi: false,
+            forced: false,
+          },
+        },
+        expected: "abk:arg",
+      },
+      {
+        source: {
+          source: {
+            content: {
+              name: "Abkhazian",
+              code2: "ab",
+              code3: "abk",
+              enabled: false,
+            },
+            hi: true,
+            forced: false,
+          },
+          target: {
+            content: {
+              name: "Aragonese",
+              code2: "an",
+              code3: "arg",
+              enabled: false,
+            },
+            hi: false,
+            forced: false,
+          },
+        },
+        expected: "abk@hi:arg",
+      },
+      {
+        source: {
+          source: {
+            content: {
+              name: "Abkhazian",
+              code2: "ab",
+              code3: "abk",
+              enabled: false,
+            },
+            hi: false,
+            forced: true,
+          },
+          target: {
+            content: {
+              name: "Aragonese",
+              code2: "an",
+              code3: "arg",
+              enabled: false,
+            },
+            hi: false,
+            forced: false,
+          },
+        },
+        expected: "abk@forced:arg",
+      },
+    ];
+
+    function testEncodeResult({ source, expected }: TestData) {
+      const encoded = encodeEqualData(source);
+
+      expect(
+        encoded,
+        `Encoded result '${encoded}' is not matched to '${expected}'`
+      ).toEqual(expected);
+    }
+
+    testValues.forEach(testEncodeResult);
+  });
+});

frontend/src/pages/Settings/Languages/equals.tsx

@@ -0,0 +1,365 @@
+import { useLanguages } from "@/apis/hooks";
+import { Action, SimpleTable } from "@/components";
+import LanguageSelector from "@/components/bazarr/LanguageSelector";
+import { languageEqualsKey } from "@/pages/Settings/keys";
+import { useFormActions } from "@/pages/Settings/utilities/FormValues";
+import { useSettingValue } from "@/pages/Settings/utilities/hooks";
+import { LOG } from "@/utilities/console";
+import { faEquals, faTrash } from "@fortawesome/free-solid-svg-icons";
+import { FontAwesomeIcon } from "@fortawesome/react-fontawesome";
+import { Button, Checkbox } from "@mantine/core";
+import { FunctionComponent, useCallback, useMemo } from "react";
+import { Column } from "react-table";
+
+interface GenericEqualTarget<T> {
+  content: T;
+  hi: boolean;
+  forced: boolean;
+}
+
+interface LanguageEqualGenericData<T> {
+  source: GenericEqualTarget<T>;
+  target: GenericEqualTarget<T>;
+}
+
+export type LanguageEqualImmediateData =
+  LanguageEqualGenericData<Language.CodeType>;
+
+export type LanguageEqualData = LanguageEqualGenericData<Language.Server>;
+
+function decodeEqualTarget(
+  text: string
+): GenericEqualTarget<Language.CodeType> | undefined {
+  const [code, decoration] = text.split("@");
+
+  if (code.length === 0) {
+    return undefined;
+  }
+
+  const forced = decoration === "forced";
+  const hi = decoration === "hi";
+
+  return {
+    content: code,
+    forced,
+    hi,
+  };
+}
+
+export function decodeEqualData(
+  text: string
+): LanguageEqualImmediateData | undefined {
+  const [first, second] = text.split(":");
+
+  const source = decodeEqualTarget(first);
+  const target = decodeEqualTarget(second);
+
+  if (source === undefined || target === undefined) {
+    return undefined;
+  }
+
+  return {
+    source,
+    target,
+  };
+}
+
+function encodeEqualTarget(data: GenericEqualTarget<Language.Server>): string {
+  let text = data.content.code3;
+  if (data.hi) {
+    text += "@hi";
+  } else if (data.forced) {
+    text += "@forced";
+  }
+
+  return text;
+}
+
+export function encodeEqualData(data: LanguageEqualData): string {
+  const source = encodeEqualTarget(data.source);
+  const target = encodeEqualTarget(data.target);
+
+  return `${source}:${target}`;
+}
+
+export function useLatestLanguageEquals(): LanguageEqualData[] {
+  const { data } = useLanguages();
+
+  const latest = useSettingValue<string[]>(languageEqualsKey);
+
+  return useMemo(
+    () =>
+      latest
+        ?.map(decodeEqualData)
+        .map((parsed) => {
+          if (parsed === undefined) {
+            return undefined;
+          }
+
+          const source = data?.find(
+            (value) => value.code3 === parsed.source.content
+          );
+          const target = data?.find(
+            (value) => value.code3 === parsed.target.content
+          );
+
+          if (source === undefined || target === undefined) {
+            return undefined;
+          }
+
+          return {
+            source: { ...parsed.source, content: source },
+            target: { ...parsed.target, content: target },
+          };
+        })
+        .filter((v): v is LanguageEqualData => v !== undefined) ?? [],
+    [data, latest]
+  );
+}
+
+interface EqualsTableProps {}
+
+const EqualsTable: FunctionComponent<EqualsTableProps> = () => {
+  const { data: languages } = useLanguages();
+  const canAdd = languages !== undefined;
+
+  const equals = useLatestLanguageEquals();
+
+  const { setValue } = useFormActions();
+
+  const setEquals = useCallback(
+    (values: LanguageEqualData[]) => {
+      const encodedValues = values.map(encodeEqualData);
+
+      LOG("info", "updating language equals data", values);
+      setValue(encodedValues, languageEqualsKey);
+    },
+    [setValue]
+  );
+
+  const add = useCallback(() => {
+    if (languages === undefined) {
+      return;
+    }
+
+    const enabled = languages.find((value) => value.enabled);
+
+    if (enabled === undefined) {
+      return;
+    }
+
+    const newValue: LanguageEqualData[] = [
+      ...equals,
+      {
+        source: {
+          content: enabled,
+          hi: false,
+          forced: false,
+        },
+        target: {
+          content: enabled,
+          hi: false,
+          forced: false,
+        },
+      },
+    ];
+
+    setEquals(newValue);
+  }, [equals, languages, setEquals]);
+
+  const update = useCallback(
+    (index: number, value: LanguageEqualData) => {
+      if (index < 0 || index >= equals.length) {
+        return;
+      }
+
+      const newValue: LanguageEqualData[] = [...equals];
+
+      newValue[index] = { ...value };
+      setEquals(newValue);
+    },
+    [equals, setEquals]
+  );
+
+  const remove = useCallback(
+    (index: number) => {
+      if (index < 0 || index >= equals.length) {
+        return;
+      }
+
+      const newValue: LanguageEqualData[] = [...equals];
+
+      newValue.splice(index, 1);
+
+      setEquals(newValue);
+    },
+    [equals, setEquals]
+  );
+
+  const columns = useMemo<Column<LanguageEqualData>[]>(
+    () => [
+      {
+        Header: "Source",
+        id: "source-lang",
+        accessor: "source",
+        Cell: ({ value: { content }, row }) => {
+          return (
+            <LanguageSelector
+              enabled
+              value={content}
+              onChange={(result) => {
+                if (result !== null) {
+                  update(row.index, {
+                    ...row.original,
+                    source: { ...row.original.source, content: result },
+                  });
+                }
+              }}
+            ></LanguageSelector>
+          );
+        },
+      },
+      {
+        id: "source-hi",
+        accessor: "source",
+        Cell: ({ value: { hi }, row }) => {
+          return (
+            <Checkbox
+              label="HI"
+              checked={hi}
+              onChange={({ currentTarget: { checked } }) => {
+                update(row.index, {
+                  ...row.original,
+                  source: {
+                    ...row.original.source,
+                    hi: checked,
+                    forced: checked ? false : row.original.source.forced,
+                  },
+                });
+              }}
+            ></Checkbox>
+          );
+        },
+      },
+      {
+        id: "source-forced",
+        accessor: "source",
+        Cell: ({ value: { forced }, row }) => {
+          return (
+            <Checkbox
+              label="Forced"
+              checked={forced}
+              onChange={({ currentTarget: { checked } }) => {
+                update(row.index, {
+                  ...row.original,
+                  source: {
+                    ...row.original.source,
+                    forced: checked,
+                    hi: checked ? false : row.original.source.hi,
+                  },
+                });
+              }}
+            ></Checkbox>
+          );
+        },
+      },
+      {
+        id: "equal-icon",
+        Cell: () => {
+          return <FontAwesomeIcon icon={faEquals} />;
+        },
+      },
+      {
+        Header: "Target",
+        id: "target-lang",
+        accessor: "target",
+        Cell: ({ value: { content }, row }) => {
+          return (
+            <LanguageSelector
+              enabled
+              value={content}
+              onChange={(result) => {
+                if (result !== null) {
+                  update(row.index, {
+                    ...row.original,
+                    target: { ...row.original.target, content: result },
+                  });
+                }
+              }}
+            ></LanguageSelector>
+          );
+        },
+      },
+      {
+        id: "target-hi",
+        accessor: "target",
+        Cell: ({ value: { hi }, row }) => {
+          return (
+            <Checkbox
+              label="HI"
+              checked={hi}
+              onChange={({ currentTarget: { checked } }) => {
+                update(row.index, {
+                  ...row.original,
+                  target: {
+                    ...row.original.target,
+                    hi: checked,
+                    forced: checked ? false : row.original.target.forced,
+                  },
+                });
+              }}
+            ></Checkbox>
+          );
+        },
+      },
+      {
+        id: "target-forced",
+        accessor: "target",
+        Cell: ({ value: { forced }, row }) => {
+          return (
+            <Checkbox
+              label="Forced"
+              checked={forced}
+              onChange={({ currentTarget: { checked } }) => {
+                update(row.index, {
+                  ...row.original,
+                  target: {
+                    ...row.original.target,
+                    forced: checked,
+                    hi: checked ? false : row.original.target.hi,
+                  },
+                });
+              }}
+            ></Checkbox>
+          );
+        },
+      },
+      {
+        id: "action",
+        accessor: "target",
+        Cell: ({ row }) => {
+          return (
+            <Action
+              label="Remove"
+              icon={faTrash}
+              color="red"
+              onClick={() => remove(row.index)}
+            ></Action>
+          );
+        },
+      },
+    ],
+    [remove, update]
+  );
+
+  return (
+    <>
+      <SimpleTable data={equals} columns={columns}></SimpleTable>
+      <Button fullWidth disabled={!canAdd} color="light" onClick={add}>
+        {canAdd ? "Add Equal" : "No Enabled Languages"}
+      </Button>
+    </>
+  );
+};
+
+export default EqualsTable;

frontend/src/pages/Settings/Languages/index.tsx

@@ -17,6 +17,7 @@ import {
 } from "../keys";
 import { useSettingValue } from "../utilities/hooks";
 import { LanguageSelector, ProfileSelector } from "./components";
+import EqualsTable from "./equals";
 import Table from "./table";
 
 export function useLatestEnabledLanguages() {
@@ -69,6 +70,13 @@ const SettingsLanguagesView: FunctionComponent = () => {
         ></LanguageSelector>
       </Section>
 
+      <Section header="Language Equals">
+        <Message>
+          Treat the following languages as equal across all providers.
+        </Message>
+        <EqualsTable></EqualsTable>
+      </Section>
+
       <Section header="Embedded Tracks Language">
         <Check
           label="Deep analyze media file to get audio tracks language."
@@ -91,7 +99,6 @@ const SettingsLanguagesView: FunctionComponent = () => {
             }}
           ></Selector>
         </CollapseBox>
-
         <Selector
           clearable
           settingKey={defaultUndEmbeddedSubtitlesLang}

frontend/src/pages/Settings/keys.ts

@@ -5,6 +5,8 @@ export const defaultUndEmbeddedSubtitlesLang =
 export const languageProfileKey = "languages-profiles";
 export const notificationsKey = "notifications-providers";
 
+export const languageEqualsKey = "settings-general-language_equals";
+
 export const pathMappingsKey = "settings-general-path_mappings";
 export const pathMappingsMovieKey = "settings-general-path_mappings_movie";
 

frontend/src/types/api.d.ts

@@ -12,6 +12,7 @@ declare namespace Language {
   type CodeType = string;
   interface Server {
     code2: CodeType;
+    code3: CodeType;
     name: string;
     enabled: boolean;
   }

frontend/src/types/settings.d.ts

@@ -25,6 +25,7 @@ interface Settings {
   titlovi: Settings.Titlovi;
   ktuvit: Settings.Ktuvit;
   notifications: Settings.Notifications;
+  language_equals: string[][];
 }
 
 declare namespace Settings {

frontend/src/utilities/languages.ts

@@ -42,3 +42,12 @@ export function useProfileItemsToLanguages(profile?: Language.Profile) {
     [data, profile?.items]
   );
 }
+
+export function useLanguageFromCode3(code3: string) {
+  const { data } = useLanguages();
+
+  return useMemo(
+    () => data?.find((value) => value.code3 === code3),
+    [data, code3]
+  );
+}

libs/subliminal_patch/core.py

@@ -153,9 +153,52 @@ class _Blacklist(list):
         return not blacklisted
 
 
+class _LanguageEquals(list):
+    """ An optional config field for the pool. It will treat a couple of languages as equal for
+    list-subtitles operations. It's optional; its methods won't do anything if an empy list
+    is set.
+
+    Example usage: [(language_instance, language_instance), ...]"""
+    def __init__(self, *args, **kwargs):
+        super().__init__(*args, **kwargs)
+
+        for item in self:
+            if len(item) != 2 or not any(isinstance(i, Language) for i in item):
+                raise ValueError(f"Not a valid equal tuple: {item}")
+
+    def check_set(self, items: set):
+        """ Check a set of languages. For example, if the set is {Language('es')} and one of the
+        equals of the instance is (Language('es'), Language('es', 'MX')), the set will now have
+        to {Language('es'), Language('es', 'MX')}.
+
+        It will return a copy of the original set to avoid messing up outside its scope.
+
+        Note that hearing_impaired and forced language attributes are not yet tested.
+        """
+        to_add = []
+        for equals in self:
+            from_, to_ = equals
+            if from_ in items:
+                logger.debug("Adding %s to %s", to_, items)
+                to_add.append(to_)
+
+        new_items = items.copy()
+        new_items.update(to_add)
+        logger.debug("New set: %s", new_items)
+        return new_items
+
+    def update_subtitle(self, subtitle):
+        for equals in self:
+            from_, to_ = equals
+            if from_ == subtitle.language:
+                logger.debug("Updating language for %s (to %s)", subtitle, to_)
+                subtitle.language = to_
+                break
+
+
 class SZProviderPool(ProviderPool):
     def __init__(self, providers=None, provider_configs=None, blacklist=None, ban_list=None, throttle_callback=None,
-                 pre_download_hook=None, post_download_hook=None, language_hook=None):
+                 pre_download_hook=None, post_download_hook=None, language_hook=None, language_equals=None):
         #: Name of providers to use
         self.providers = set(providers or [])
 
@@ -170,6 +213,8 @@ class SZProviderPool(ProviderPool):
         #: Should be a dict of 2 lists of strings
         self.ban_list = _Banlist(**(ban_list or {'must_contain': [], 'must_not_contain': []}))
 
+        self.lang_equals = _LanguageEquals(language_equals or [])
+
         self.throttle_callback = throttle_callback
 
         self.pre_download_hook = pre_download_hook
@@ -185,7 +230,7 @@ class SZProviderPool(ProviderPool):
         self.provider_configs = _ProviderConfigs(self)
         self.provider_configs.update(provider_configs or {})
 
-    def update(self, providers, provider_configs, blacklist, ban_list):
+    def update(self, providers, provider_configs, blacklist, ban_list, language_equals=None):
         # Check if the pool was initialized enough hours ago
         self._check_lifetime()
 
@@ -222,6 +267,7 @@ class SZProviderPool(ProviderPool):
 
         self.blacklist = _Blacklist(blacklist or [])
         self.ban_list = _Banlist(**ban_list or {'must_contain': [], 'must_not_contain': []})
+        self.lang_equals = _LanguageEquals(language_equals or [])
 
         return updated
 
@@ -299,7 +345,7 @@ class SZProviderPool(ProviderPool):
             return []
 
         # check supported languages
-        provider_languages = provider_registry[provider].languages & use_languages
+        provider_languages = self.lang_equals.check_set(set(provider_registry[provider].languages)) & use_languages
         if not provider_languages:
             logger.info('Skipping provider %r: no language to search for', provider)
             return []
@@ -312,6 +358,8 @@ class SZProviderPool(ProviderPool):
             seen = []
             out = []
             for s in results:
+                self.lang_equals.update_subtitle(s)
+
                 if not self.blacklist.is_valid(provider, s):
                     continue
 
@@ -569,7 +617,7 @@ class SZProviderPool(ProviderPool):
                 continue
 
             # add the languages for this provider
-            languages.append({'provider': name, 'languages': provider_languages})
+            languages.append({'provider': name, 'languages': self.lang_equals.check_set(set(provider_languages))})
 
         return languages
 

tests/bazarr/app/test_get_providers.py

@@ -1,45 +0,0 @@
-import pytest
-
-import inspect
-
-from bazarr.app import get_providers
-
-
-def test_get_providers_auth():
-    for val in get_providers.get_providers_auth().values():
-        assert isinstance(val, dict)
-
-
-def test_get_providers_auth_with_provider_registry():
-    """Make sure all providers will be properly initialized with bazarr
-    configs"""
-    from subliminal_patch.extensions import provider_registry
-
-    auths = get_providers.get_providers_auth()
-    for key, val in auths.items():
-        provider = provider_registry[key]
-        sign = inspect.signature(provider.__init__)
-        for sub_key in val.keys():
-            if sub_key not in sign.parameters:
-                raise ValueError(f"'{sub_key}' parameter not present in {provider}")
-
-            assert sign.parameters[sub_key] is not None
-
-
-def test_get_providers_auth_embeddedsubtitles():
-    item = get_providers.get_providers_auth()["embeddedsubtitles"]
-    assert isinstance(item["included_codecs"], list)
-    assert isinstance(item["hi_fallback"], bool)
-    assert isinstance(item["cache_dir"], str)
-    assert isinstance(item["ffprobe_path"], str)
-    assert isinstance(item["ffmpeg_path"], str)
-    assert isinstance(item["timeout"], str)
-    assert isinstance(item["unknown_as_english"], bool)
-
-
-def test_get_providers_auth_karagarga():
-    item = get_providers.get_providers_auth()["karagarga"]
-    assert item["username"] is not None
-    assert item["password"] is not None
-    assert item["f_username"] is not None
-    assert item["f_password"] is not None

tests/bazarr/conftest.py

@@ -3,5 +3,6 @@ import logging
 
 os.environ["NO_CLI"] = "true"
 os.environ["SZ_USER_AGENT"] = "test"
+os.environ["BAZARR_VERSION"] = "test" # fixme
 
 logging.getLogger("rebulk").setLevel(logging.WARNING)

tests/bazarr/test_app_get_providers.py

@@ -0,0 +1,115 @@
+import inspect
+
+import pytest
+from subliminal_patch.core import Language
+
+from bazarr.app import get_providers
+
+
+def test_get_providers_auth():
+    for val in get_providers.get_providers_auth().values():
+        assert isinstance(val, dict)
+
+
+def test_get_providers_auth_with_provider_registry():
+    """Make sure all providers will be properly initialized with bazarr
+    configs"""
+    from subliminal_patch.extensions import provider_registry
+
+    auths = get_providers.get_providers_auth()
+    for key, val in auths.items():
+        provider = provider_registry[key]
+        sign = inspect.signature(provider.__init__)
+        for sub_key in val.keys():
+            if sub_key not in sign.parameters:
+                raise ValueError(f"'{sub_key}' parameter not present in {provider}")
+
+            assert sign.parameters[sub_key] is not None
+
+
+def test_get_providers_auth_embeddedsubtitles():
+    item = get_providers.get_providers_auth()["embeddedsubtitles"]
+    assert isinstance(item["included_codecs"], list)
+    assert isinstance(item["hi_fallback"], bool)
+    assert isinstance(item["cache_dir"], str)
+    assert isinstance(item["ffprobe_path"], str)
+    assert isinstance(item["ffmpeg_path"], str)
+    assert isinstance(item["timeout"], str)
+    assert isinstance(item["unknown_as_english"], bool)
+
+
+def test_get_providers_auth_karagarga():
+    item = get_providers.get_providers_auth()["karagarga"]
+    assert item["username"] is not None
+    assert item["password"] is not None
+    assert item["f_username"] is not None
+    assert item["f_password"] is not None
+
+
+def test_get_language_equals_default_settings():
+    assert isinstance(get_providers.get_language_equals(), list)
+
+
+def test_get_language_equals_injected_settings_invalid():
+    config = get_providers.settings
+    config.set("general", "language_equals", '["invalid"]')
+    assert not get_providers.get_language_equals(config)
+
+
+def test_get_language_equals_injected_settings_valid():
+    config = get_providers.settings
+    config.set("general", "language_equals", '["spa:spa-MX"]')
+
+    result = get_providers.get_language_equals(config)
+    assert result == [(Language("spa"), Language("spa", "MX"))]
+
+
+@pytest.mark.parametrize(
+    "config_value,expected",
+    [
+        ('["spa:spl"]', (Language("spa"), Language("spa", "MX"))),
+        ('["por:pob"]', (Language("por"), Language("por", "BR"))),
+        ('["zho:zht"]', (Language("zho"), Language("zho", "TW"))),
+    ],
+)
+def test_get_language_equals_injected_settings_custom_lang_alpha3(
+    config_value, expected
+):
+    config = get_providers.settings
+
+    config.set("general", "language_equals", config_value)
+
+    result = get_providers.get_language_equals(config)
+    assert result == [expected]
+
+
+def test_get_language_equals_injected_settings_multiple():
+    config = get_providers.settings
+
+    config.set(
+        "general",
+        "language_equals",
+        "['eng@hi:eng', 'spa:spl', 'spa@hi:spl', 'spl@hi:spl']",
+    )
+
+    result = get_providers.get_language_equals(config)
+    assert len(result) == 4
+
+
+def test_get_language_equals_injected_settings_valid_multiple():
+    config = get_providers.settings
+    config.set("general", "language_equals", '["spa:spa-MX", "spa-MX:spa"]')
+
+    result = get_providers.get_language_equals(config)
+    assert result == [
+        (Language("spa"), Language("spa", "MX")),
+        (Language("spa", "MX"), Language("spa")),
+    ]
+
+
+def test_get_language_equals_injected_settings_hi():
+    config = get_providers.settings
+    config.set("general", "language_equals", '["eng@hi:eng"]')
+
+    result = get_providers.get_language_equals(config)
+    assert result == [(Language("eng", hi=True), Language("eng"))]

tests/bazarr/test_subtitles_pool.py

@@ -0,0 +1,10 @@
+from bazarr.subtitles import pool
+
+
+def test_init_pool():
+    assert pool._init_pool("movie")
+
+
+def test_pool_update():
+    pool_ = pool._init_pool("movie")
+    assert pool._pool_update(pool_, "movie")

tests/subliminal_patch/test_core.py

@@ -70,3 +70,99 @@ def test_pool_update_discarded_providers_2(pool_instance):
 
     # Provider should not disappear from discarded providers
     assert pool_instance.discarded_providers == {"argenteam"}
+
+
+def test_language_equals_init():
+    assert core._LanguageEquals([(core.Language("spa"), core.Language("spa", "MX"))])
+
+
+def test_language_equals_init_invalid():
+    with pytest.raises(ValueError):
+        assert core._LanguageEquals([(core.Language("spa", "MX"),)])
+
+
+def test_language_equals_init_empty_list_gracefully():
+    assert core._LanguageEquals([]) == []
+
+
+@pytest.mark.parametrize(
+    "langs",
+    [
+        [(core.Language("spa"), core.Language("spa", "MX"))],
+        [(core.Language("por"), core.Language("por", "BR"))],
+        [(core.Language("zho"), core.Language("zho", "TW"))],
+    ],
+)
+def test_language_equals_check_set(langs):
+    equals = core._LanguageEquals(langs)
+    lang_set = {langs[0]}
+    assert equals.check_set(lang_set) == set(langs)
+
+
+def test_language_equals_check_set_do_nothing():
+    equals = core._LanguageEquals([(core.Language("eng"), core.Language("spa"))])
+    lang_set = {core.Language("spa")}
+    assert equals.check_set(lang_set) == {core.Language("spa")}
+
+
+def test_language_equals_check_set_do_nothing_w_forced():
+    equals = core._LanguageEquals(
+        [(core.Language("spa", forced=True), core.Language("spa", "MX"))]
+    )
+    lang_set = {core.Language("spa")}
+    assert equals.check_set(lang_set) == {core.Language("spa")}
+
+
+@pytest.fixture
+def language_equals_pool_intance():
+    equals = [(core.Language("spa"), core.Language("spa", "MX"))]
+    yield core.SZProviderPool({"subdivx"}, language_equals=equals)
+
+
+def test_language_equals_pool_intance_list_subtitles(
+    language_equals_pool_intance, movies
+):
+    subs = language_equals_pool_intance.list_subtitles(
+        movies["dune"], {core.Language("spa")}
+    )
+    assert subs
+    assert all(sub.language == core.Language("spa", "MX") for sub in subs)
+
+
+def test_language_equals_pool_intance_list_subtitles_reversed(movies):
+    equals = [(core.Language("spa", "MX"), core.Language("spa"))]
+    language_equals_pool_intance = core.SZProviderPool(
+        {"subdivx"}, language_equals=equals
+    )
+    subs = language_equals_pool_intance.list_subtitles(
+        movies["dune"], {core.Language("spa")}
+    )
+    assert subs
+    assert all(sub.language == core.Language("spa") for sub in subs)
+
+
+def test_language_equals_pool_intance_list_subtitles_empty_lang_equals(movies):
+    language_equals_pool_intance = core.SZProviderPool(
+        {"subdivx"}, language_equals=None
+    )
+    subs = language_equals_pool_intance.list_subtitles(
+        movies["dune"], {core.Language("spa")}
+    )
+    assert subs
+    assert not all(sub.language == core.Language("spa", "MX") for sub in subs)
+
+
+def test_language_equals_pool_intance_list_subtitles_return_nothing(movies):
+    equals = [
+        (core.Language("spa", "MX"), core.Language("eng")),
+        (core.Language("spa"), core.Language("eng")),
+    ]
+    language_equals_pool_intance = core.SZProviderPool(
+        {"subdivx"}, language_equals=equals
+    )
+    subs = language_equals_pool_intance.list_subtitles(
+        movies["dune"], {core.Language("spa")}
+    )
+    assert not language_equals_pool_intance.download_best_subtitles(
+        subs, movies["dune"], {core.Language("spa")}
+    )