entari-plugin-hyw 4.0.0rc3__py3-none-any.whl → 4.0.0rc4__py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Potentially problematic release.
This version of entari-plugin-hyw might be problematic. Click here for more details.
- entari_plugin_hyw/pipeline.py +1 -22
- {entari_plugin_hyw-4.0.0rc3.dist-info → entari_plugin_hyw-4.0.0rc4.dist-info}/METADATA +1 -1
- {entari_plugin_hyw-4.0.0rc3.dist-info → entari_plugin_hyw-4.0.0rc4.dist-info}/RECORD +5 -5
- {entari_plugin_hyw-4.0.0rc3.dist-info → entari_plugin_hyw-4.0.0rc4.dist-info}/WHEEL +0 -0
- {entari_plugin_hyw-4.0.0rc3.dist-info → entari_plugin_hyw-4.0.0rc4.dist-info}/top_level.txt +0 -0
entari_plugin_hyw/pipeline.py
CHANGED
|
@@ -196,17 +196,8 @@ class ProcessingPipeline:
|
|
|
196
196
|
|
|
197
197
|
# Deduplicate while preserving order and filter blocked domains
|
|
198
198
|
final_fetch_urls = []
|
|
199
|
-
blocked_domains = getattr(self.config, "fetch_blocked_domains", None)
|
|
200
|
-
if blocked_domains is None:
|
|
201
|
-
blocked_domains = ["wikipedia.org", "csdn.net", "sohu.com", "sogou.com"]
|
|
202
|
-
if isinstance(blocked_domains, str):
|
|
203
|
-
blocked_domains = [d.strip() for d in blocked_domains.split(",")]
|
|
204
|
-
|
|
205
199
|
for url in raw_fetch_urls:
|
|
206
200
|
if url and url not in final_fetch_urls:
|
|
207
|
-
# Check blocklist
|
|
208
|
-
if any(domain in url.lower() for domain in blocked_domains):
|
|
209
|
-
continue
|
|
210
201
|
final_fetch_urls.append(url)
|
|
211
202
|
|
|
212
203
|
fetch_urls = final_fetch_urls
|
|
@@ -574,19 +565,7 @@ class ProcessingPipeline:
|
|
|
574
565
|
self._search_error = str(e)
|
|
575
566
|
raise e
|
|
576
567
|
|
|
577
|
-
# Filter blocked domains
|
|
578
|
-
blocked_domains = getattr(self.config, "fetch_blocked_domains", ["wikipedia.org", "csdn.net", "baidu.com"])
|
|
579
|
-
if isinstance(blocked_domains, str):
|
|
580
|
-
blocked_domains = [d.strip() for d in blocked_domains.split(",")]
|
|
581
|
-
|
|
582
|
-
# Use list comprehension for filtering
|
|
583
|
-
original_count = len(web)
|
|
584
|
-
web = [
|
|
585
|
-
item for item in web
|
|
586
|
-
if not any(blocked in item.get("url", "").lower() for blocked in blocked_domains)
|
|
587
|
-
]
|
|
588
|
-
if len(web) < original_count:
|
|
589
|
-
logger.info(f"Filtered {original_count - len(web)} blocked search results.")
|
|
568
|
+
# Filter blocked domains removed per user request (handled in search query)
|
|
590
569
|
|
|
591
570
|
# Cache results and assign global IDs
|
|
592
571
|
for item in web:
|
|
@@ -2,7 +2,7 @@ entari_plugin_hyw/__init__.py,sha256=XRjnYDWX-dfMi33yD1CcfejHcQyySXTrQVlo95dVfis
|
|
|
2
2
|
entari_plugin_hyw/history.py,sha256=zYtON0FgkA_AcXerLV335OzpIP30eAxDEp7NHCFFXis,7016
|
|
3
3
|
entari_plugin_hyw/image_cache.py,sha256=t8pr1kgH2ngK9IhrBAhzUqhBWERNztUywMzgCFZEtQk,9899
|
|
4
4
|
entari_plugin_hyw/misc.py,sha256=b0KPCUrf1GuCY0Mx29Sn5sREg1fxR1OyYEl4O4Omsys,4553
|
|
5
|
-
entari_plugin_hyw/pipeline.py,sha256=
|
|
5
|
+
entari_plugin_hyw/pipeline.py,sha256=CFp8D9OAJ3nfDsJwKNHb42ZLHYoUL5k9wykrcWC_Tno,52501
|
|
6
6
|
entari_plugin_hyw/prompts.py,sha256=BqX8MNdBG7M6_oRAe1jCiSCgcbRchFAoWpgF9WB3-0o,2348
|
|
7
7
|
entari_plugin_hyw/render_vue.py,sha256=cdIsvuGIC-01Nmc9vgYnGQIY5CGh_ZdqphAhFP4h0Qo,12426
|
|
8
8
|
entari_plugin_hyw/search.py,sha256=bN958sSXYIz7fMmH_ZoS1777c1wXatV05-rvfEq_YHQ,30383
|
|
@@ -82,7 +82,7 @@ entari_plugin_hyw/card-ui/src/components/HelloWorld.vue,sha256=yvBIzJua9BfikUOR1
|
|
|
82
82
|
entari_plugin_hyw/card-ui/src/components/MarkdownContent.vue,sha256=JO1sKThQB42bVl5382BkxpDEyRpAK7oGcmdj1LY5P64,13190
|
|
83
83
|
entari_plugin_hyw/card-ui/src/components/SectionCard.vue,sha256=owcDNx2JYVmF2J5SYCroR2gvg_cPApQsNunjK1WJpVI,1433
|
|
84
84
|
entari_plugin_hyw/card-ui/src/components/StageCard.vue,sha256=cC0TeJ4AhtN3sJEXenVpl2Eq5901GKC59qpDUNH2SZE,10964
|
|
85
|
-
entari_plugin_hyw-4.0.
|
|
86
|
-
entari_plugin_hyw-4.0.
|
|
87
|
-
entari_plugin_hyw-4.0.
|
|
88
|
-
entari_plugin_hyw-4.0.
|
|
85
|
+
entari_plugin_hyw-4.0.0rc4.dist-info/METADATA,sha256=0Ytj2E6tpljuG281p44mEEusfc3V-ZUB1V5rvEswwos,3746
|
|
86
|
+
entari_plugin_hyw-4.0.0rc4.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
|
|
87
|
+
entari_plugin_hyw-4.0.0rc4.dist-info/top_level.txt,sha256=TIDsn6XPs6KA5e3ezsE65JoXsy03ejDdrB41I4SPjmo,18
|
|
88
|
+
entari_plugin_hyw-4.0.0rc4.dist-info/RECORD,,
|
|
File without changes
|
|
File without changes
|