[mod] engine: MyMemory Translated - typification and html to text (#6132)

The implementation is normalized, type annotations are applied, and the results
are freed from the HTML markup (which is partially present).

Signed-off-by: Markus Heiser <markus.heiser@darmarit.de>

searx/engines/translated.py

@@ -1,56 +1,73 @@
 # SPDX-License-Identifier: AGPL-3.0-or-later
 """MyMemory Translated"""
 
+import typing as t
+
 import urllib.parse
 
+from searx.utils import html_to_text
 from searx.result_types import EngineResults
 
+if t.TYPE_CHECKING:
+    from searx.extended_types import SXNG_Response
+    from searx.search.processors import OnlineDictParams
+#
 # about
 about = {
-    "website": 'https://mymemory.translated.net/',
+    "website": "https://mymemory.translated.net/",
     "wikidata_id": None,
-    "official_api_documentation": 'https://mymemory.translated.net/doc/spec.php',
+    "official_api_documentation": "https://mymemory.translated.net/doc/spec.php",
     "use_official_api": True,
     "require_api_key": False,
-    "results": 'JSON',
+    "results": "JSON",
 }
 
-engine_type = 'online_dictionary'
+engine_type = "online_dictionary"
-categories = ['general', 'translate']
+categories = ["general", "translate"]
 api_url = "https://api.mymemory.translated.net"
 web_url = "https://mymemory.translated.net"
 weight = 100
 
-api_key = ''
+api_key = ""
 
 
-def request(query, params):  # pylint: disable=unused-argument
+def request(_: str, params: "OnlineDictParams") -> None:
 
-    args = {"q": params["query"], "langpair": f"{params['from_lang'][1]}|{params['to_lang'][1]}"}
+    args = {
+        "q": params["query"],
+        "langpair": f"{params['from_lang'][1]}|{params['to_lang'][1]}",
+    }
     if api_key:
         args["key"] = api_key
 
     params['url'] = f"{api_url}/get?{urllib.parse.urlencode(args)}"
-    return params
 
 
-def response(resp) -> EngineResults:
+def response(resp: "SXNG_Response") -> EngineResults:
+
     results = EngineResults()
-    data = resp.json()
+    data: dict[str, t.Any] = resp.json()
+    params: "OnlineDictParams" = resp.search_params  # pyright: ignore[reportAssignmentType]
 
     args = {
-        "q": resp.search_params["query"],
+        "q": params["query"],
-        "lang": resp.search_params.get("searxng_locale", "en"),  # ui language
+        "lang": params.get("searxng_locale", "en"),  # ui language
-        "sl": resp.search_params['from_lang'][1],
+        "sl": params["from_lang"][1],
-        "tl": resp.search_params['to_lang'][1],
+        "tl": params["to_lang"][1],
     }
 
     link = f"{web_url}/search.php?{urllib.parse.urlencode(args)}"
-    text = data['responseData']['translatedText']
+    text: str = html_to_text(data["responseData"]["translatedText"])
 
-    examples = [f"{m['segment']} : {m['translation']}" for m in data['matches'] if m['translation'] != text]
+    examples: set[str] = set()
+    match: dict[str, str]
+    for match in data["matches"]:
+        _text = html_to_text(match["translation"])
+        if _text != text:
+            _seg = html_to_text(match["segment"])
+            examples.add(f"{_seg} : {_text}")
 
-    item = results.types.Translations.Item(text=text, examples=examples)
+    item = results.types.Translations.Item(text=text, examples=list(examples))
     results.add(results.types.Translations(translations=[item], url=link))
 
     return results

searx/extended_types.py

@@ -30,7 +30,7 @@ import httpx
 if typing.TYPE_CHECKING:
     import searx.preferences
     import searx.results
-    from searx.search.processors import OnlineParamTypes
+    from searx.search.processors import OnlineParamTypes, OnlineDictParams, OnlineCurrenciesParams
 
 
 class SXNG_Request(flask.Request):
@@ -83,4 +83,4 @@ class SXNG_Response(httpx.Response):
     """
 
     ok: bool
-    search_params: "OnlineParamTypes"
+    search_params: "OnlineParamTypes | OnlineDictParams | OnlineCurrenciesParams"