feat(agent): scrape linked pages to ground answers in real content

Adds `_fetch_page_content()`, which downloads a URL with a
browser-like User-Agent and uses BeautifulSoup with the lxml
parser to strip scripts, navigation, and other non-content
markup, returning up to 3000 characters of readable text.

The response validator now also compares generated replies
against this scraped text, flagging named entities and claims
that do not appear in the retrieved page content.

Dependencies: adds beautifulsoup4==4.12.3 and pins lxml==5.3.0
for parser compatibility.
This commit is contained in:
2026-10-26 16:48:29 -03:00
parent 9eed02645b
commit 723ca1506b
231 changed files with 20508 additions and 5330 deletions
@@ -0,0 +1,79 @@
beautifulsoup4-4.12.3.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
beautifulsoup4-4.12.3.dist-info/METADATA,sha256=UkOS1koIjlakIy9Q1u2yCNwDEFOUZSrLcsbV-mTInz4,3790
beautifulsoup4-4.12.3.dist-info/RECORD,,
beautifulsoup4-4.12.3.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
beautifulsoup4-4.12.3.dist-info/WHEEL,sha256=mRYSEL3Ih6g5a_CVMIcwiF__0Ae4_gLYh01YFNwiq1k,87
beautifulsoup4-4.12.3.dist-info/licenses/AUTHORS,sha256=uSIdbrBb1sobdXl7VrlUvuvim2dN9kF3MH4Edn0WKGE,2176
beautifulsoup4-4.12.3.dist-info/licenses/LICENSE,sha256=VbTY1LHlvIbRDvrJG3TIe8t3UmsPW57a-LnNKtxzl7I,1441
bs4/__init__.py,sha256=kq32cCtQiNjjU9XwjD0b1jdXN5WEC87nJqSSW3PhVkM,33822
bs4/__pycache__/__init__.cpython-311.pyc,,
bs4/__pycache__/css.cpython-311.pyc,,
bs4/__pycache__/dammit.cpython-311.pyc,,
bs4/__pycache__/diagnose.cpython-311.pyc,,
bs4/__pycache__/element.cpython-311.pyc,,
bs4/__pycache__/formatter.cpython-311.pyc,,
bs4/builder/__init__.py,sha256=nwb35ftjcwzOs2WkjVm1zvfi7FxSyJP-nN1YheIVT14,24566
bs4/builder/__pycache__/__init__.cpython-311.pyc,,
bs4/builder/__pycache__/_html5lib.cpython-311.pyc,,
bs4/builder/__pycache__/_htmlparser.cpython-311.pyc,,
bs4/builder/__pycache__/_lxml.cpython-311.pyc,,
bs4/builder/_html5lib.py,sha256=0w-hmPM5wWR2iDuRCR6MvY6ZPXbg_hgddym-YWqj03s,19114
bs4/builder/_htmlparser.py,sha256=_VD5Z08j6A9YYMR4y7ZTfdMzwiCBsSUQAPuHiYB-WZI,14923
bs4/builder/_lxml.py,sha256=yKdMx1kdX7H2CopwSWEYm4Sgrfkd-WDj8HbskcaLauU,14948
bs4/css.py,sha256=gqGaHRrKeCRF3gDqxzeU0uclOCeSsTpuW9gUaSnJeWc,10077
bs4/dammit.py,sha256=G0cQfsEqfwJ-FIQMkXgCJwSHMn7t9vPepCrud6fZEKk,41158
bs4/diagnose.py,sha256=uAwdDugL_67tB-BIwDIFLFbiuzGxP2wQzJJ4_bGYUrA,7195
bs4/element.py,sha256=Dsol2iehkSjk10GzYgwFyjUEgpqmYZpyaAmbL0rWM2w,92845
bs4/formatter.py,sha256=Bu4utAQYT9XDJaPPpTRM-dyxJDVLdxf_as-IU5gSY8A,7188
bs4/tests/__init__.py,sha256=NydTegds_r7MoOEuQLS6TFmTA9TwK3KxJhwEkqjCGTQ,48392
bs4/tests/__pycache__/__init__.cpython-311.pyc,,
bs4/tests/__pycache__/test_builder.cpython-311.pyc,,
bs4/tests/__pycache__/test_builder_registry.cpython-311.pyc,,
bs4/tests/__pycache__/test_css.cpython-311.pyc,,
bs4/tests/__pycache__/test_dammit.cpython-311.pyc,,
bs4/tests/__pycache__/test_docs.cpython-311.pyc,,
bs4/tests/__pycache__/test_element.cpython-311.pyc,,
bs4/tests/__pycache__/test_formatter.cpython-311.pyc,,
bs4/tests/__pycache__/test_fuzz.cpython-311.pyc,,
bs4/tests/__pycache__/test_html5lib.cpython-311.pyc,,
bs4/tests/__pycache__/test_htmlparser.cpython-311.pyc,,
bs4/tests/__pycache__/test_lxml.cpython-311.pyc,,
bs4/tests/__pycache__/test_navigablestring.cpython-311.pyc,,
bs4/tests/__pycache__/test_pageelement.cpython-311.pyc,,
bs4/tests/__pycache__/test_soup.cpython-311.pyc,,
bs4/tests/__pycache__/test_tag.cpython-311.pyc,,
bs4/tests/__pycache__/test_tree.cpython-311.pyc,,
bs4/tests/fuzz/clusterfuzz-testcase-minimized-bs4_fuzzer-4670634698080256.testcase,sha256=yUdXkbpNK7LVOQ0LBHMoqZ1rWaBfSXWytoO_xdSm7Ho,15
bs4/tests/fuzz/clusterfuzz-testcase-minimized-bs4_fuzzer-4818336571064320.testcase,sha256=Uv_dx4a43TSfoNkjU-jHW2nSXkqHFg4XdAw7SWVObUk,23
bs4/tests/fuzz/clusterfuzz-testcase-minimized-bs4_fuzzer-4999465949331456.testcase,sha256=OEyVA0Ej4FxswOElrUNt0In4s4YhrmtaxE_NHGZvGtg,30
bs4/tests/fuzz/clusterfuzz-testcase-minimized-bs4_fuzzer-5000587759190016.testcase,sha256=G4vpNBOz-RwMpi6ewEgNEa13zX0sXhmL7VHOyIcdKVQ,15347
bs4/tests/fuzz/clusterfuzz-testcase-minimized-bs4_fuzzer-5167584867909632.testcase,sha256=3d8z65o4p7Rur-RmCHoOjzqaYQ8EAtjmiBYTHNyAdl4,19469
bs4/tests/fuzz/clusterfuzz-testcase-minimized-bs4_fuzzer-5270998950477824.testcase,sha256=NfGIlit1k40Ip3mlnBkYOkIDJX6gHtjlErwl7gsBjAQ,12
bs4/tests/fuzz/clusterfuzz-testcase-minimized-bs4_fuzzer-5375146639360000.testcase,sha256=xy4i1U0nhFHcnyc5pRKS6JRMvuoCNUur-Scor6UxIGw,4317
bs4/tests/fuzz/clusterfuzz-testcase-minimized-bs4_fuzzer-5492400320282624.testcase,sha256=Q-UTYpQBUsWoMgIUspUlzveSI-41s4ABC3jajRb-K0o,11502
bs4/tests/fuzz/clusterfuzz-testcase-minimized-bs4_fuzzer-5703933063462912.testcase,sha256=2bq3S8KxZgk8EajLReHD8m4_0Lj_nrkyJAxB_z_U0D0,5
bs4/tests/fuzz/clusterfuzz-testcase-minimized-bs4_fuzzer-5843991618256896.testcase,sha256=MZDu31LPLfgu6jP9IZkrlwNes3f_sL8WFP5BChkUKdY,35
bs4/tests/fuzz/clusterfuzz-testcase-minimized-bs4_fuzzer-5984173902397440.testcase,sha256=w58r-s6besG5JwPXpnz37W2YTj9-_qxFbk6hiEnKeIQ,51495
bs4/tests/fuzz/clusterfuzz-testcase-minimized-bs4_fuzzer-6124268085182464.testcase,sha256=q8rkdMECEXKcqVhOf5zWHkSBTQeOPt0JiLg2TZiPCuk,10380
bs4/tests/fuzz/clusterfuzz-testcase-minimized-bs4_fuzzer-6241471367348224.testcase,sha256=QfzoOxKwNuqG-4xIrea6MOQLXhfAAOQJ0r9u-J6kSNs,19
bs4/tests/fuzz/clusterfuzz-testcase-minimized-bs4_fuzzer-6306874195312640.testcase,sha256=MJ2pHFuuCQUiQz1Kor2sof7LWeRERQ6QK43YNqQHg9o,47
bs4/tests/fuzz/clusterfuzz-testcase-minimized-bs4_fuzzer-6450958476902400.testcase,sha256=EItOpSdeD4ewK-qgJ9vtxennwn_huguzXgctrUT7fqE,3546
bs4/tests/fuzz/clusterfuzz-testcase-minimized-bs4_fuzzer-6600557255327744.testcase,sha256=a2aJTG4FceGSJXsjtxoS8S4jk_8rZsS3aznLkeO2_dY,124
bs4/tests/fuzz/crash-0d306a50c8ed8bcd0785b67000fcd5dea1d33f08.testcase,sha256=jRFRtCKlP3-3EDLc_iVRTcE6JNymv0rYcVM6qRaPrxI,2607
bs4/tests/fuzz/crash-ffbdfa8a2b26f13537b68d3794b0478a4090ee4a.testcase,sha256=7NsdCiXWAhNkmoW1pvF7rbZExyLAQIWtDtSHXIsH6YU,103
bs4/tests/test_builder.py,sha256=nc2JE5EMrEf-p24qhf2R8qAV5PpFiOuNpYCmtmCjlTI,1115
bs4/tests/test_builder_registry.py,sha256=7WLj2prjSHGphebnrjQuI6JYr03Uy_c9_CkaFSQ9HRo,5114
bs4/tests/test_css.py,sha256=jCcgIWem3lyPa5AjhAk9S6fWI07hk1rg0v8coD7bEtI,17279
bs4/tests/test_dammit.py,sha256=MbSmRN6VEP0Rm56-w6Ja0TW8eC-8ZxOJ-wXWVf_hRi8,15451
bs4/tests/test_docs.py,sha256=xoAxnUfoQ7aRqGImwW_9BJDU8WNMZHIuvWqVepvWXt8,1127
bs4/tests/test_element.py,sha256=92oRSRoGk8gIXAbAGHErKzocx2MK32TqcQdUJ-dGQMo,2377
bs4/tests/test_formatter.py,sha256=eTzj91Lmhv90z-WiHjK3sBJZm0hRk0crFY1TZaXstCY,4148
bs4/tests/test_fuzz.py,sha256=_K2utiYVkZ22mvh03g8CBioFU1QDJaff1vTaDyXhxNk,6972
bs4/tests/test_html5lib.py,sha256=2-ipm-_MaPt37WTxEd5DodUTNhS4EbLFKPRaO6XSCW4,8322
bs4/tests/test_htmlparser.py,sha256=wnngcIlzjEwH21JFfu_mgt6JdpLt0ncJfLcGT7HeGw0,6256
bs4/tests/test_lxml.py,sha256=nQCmLt7bWk0id7xMumZw--PzEe1xF9PTQn3lvHyNC6I,7635
bs4/tests/test_navigablestring.py,sha256=RGSgziNf7cZnYdEPsoqL1B2I68TUJp1JmEQVxbh_ryA,5081
bs4/tests/test_pageelement.py,sha256=VdGjUxx3RhjqmNsJ92ao6VZC_YD7T8mdLkDZjosOYeE,14274
bs4/tests/test_soup.py,sha256=JmnAPLE1_GXm0wmwEUN7icdvBz9HDch-qoU2mT_TDrs,19877
bs4/tests/test_tag.py,sha256=FBPDUisDCbFmvl5HmTtN49CGo3YoUXh5Wiuw5FMLS5E,9616
bs4/tests/test_tree.py,sha256=n9nTQOzJb3-ZnZ6AkmMdZQ5TYcTUPnqHoVgal0mYXfg,48129