/

/

sber-deep-context

Обзор ДокументацияВойти

/

/

sber-deep-context

Аналитика

Безопасность

sber-deep-context

Описание

Решение кейса "AI-помощник по бизнес-контексту" в рамках AI-хакатона 2026 от команды Digital Reporting Сбера

77.84 MiB

В избранном0

Языки

Python67,3%
HTML20,6%
CSS12,1%

Сообщить о нарушении

Коммиты:

A

ari_ira17

docs: add main screenshot to README to showcase the app instantly

14 сен 2026, 09:48

14 сен 2026, 09:4895a068a

fix: pass sandbox documents to streaming astream_ask pipeline

20 дней назад

fix: pass sandbox documents to streaming astream_ask pipeline

20 дней назад

fix(rag): map missing metadata to DocumentContext and lower min_reranker_score to fix empty search results

20 дней назад

feat(data): clean dataset null attachments and add LanceDB ingestion script

20 дней назад

chore: remove deprecated docs, examples, and test files

19 дней назад

refactor: remove unused tests and strip comments from python files

20 дней назад

refactor: remove unused tests and strip comments from python files

20 дней назад

feat: add parsers

21 день назад

refactor: remove unused tests and strip comments from python files

20 дней назад

feat(agents): implement adaptive model routing (Fast-Path Lite vs Deep Max) and cache RAG engine

20 дней назад

refactor: remove unused tests and strip comments from python files

20 дней назад

docs: add environment template and knowledge base archive

19 дней назад

fix(export): fix docx export, collapse adjacent citations, and clean up exported file headers

20 дней назад

chore: remove Docker support and restore local instructions

20 дней назад

docs: add main screenshot to README to showcase the app instantly

19 дней назад

docs: rename ui_showcase to SCREENSHOTS.md for better visibility

19 дней назад

meridian_hackathon_knowledge_base.zip

docs: add environment template and knowledge base archive

19 дней назад

requirements.txt

fix: remove non-existent fastmcp dependency from requirements

20 дней назад

README.md