Commit History
-
Switch task model and add max tokens limits (#1036)
Unverified signatureSwitch task modle and add max tokens limits86bc2ea
-
Add websearch controls for assistants (#812)
Unverified signature* remove query modifiers from generateQuery * Add backend for assistant RAG * Add front-end for updating RAG assistant * enable web parser to return plaintext directly for matching headers * Update websearch flow for handling assistant rag preferences * Add our old blocklist to .env.template * Enable websearch to run on messages depending on assistant requirements * reorganized imports * Rename vars * use projection * Add environment variable for assistant rag * fix assistant rag on runwebsearch * fix styling if rag is disabled * make sure we always omit credentials when fetching web pages * Add new checks for SSRF, with a new env var `ENABLE_LOCAL_FETCH` * Use DNS to check if the links are local or not * Add a websearch indicator * Add more tags to parser * Add indicators * Display RAG options in settings view * ui * fix rag detection * bit more spacing * fix button position in assistant form * wording (mainly) …3f5871c
-
[Fix] small fixes in websearch generateQuery (#727)
MishigUnverified signature13489e8
-
Use JSON5 for parsing all config vars (#671)
Unverified signaturef09c5f3
-
Add allow & blocklist for websearch (#651)
Unverified signature* Add allow & blocklist for websearch * fix start whitespacebdb1788
-
Modular backends & support for openAI & AWS endpoints (#541)
Unverified signature* Fix the response Signed-off-by: Hung-Han (Henry) Chen <chenhungh@gmail.com> * Should use /completions Signed-off-by: Hung-Han (Henry) Chen <chenhungh@gmail.com> * Use async generator Signed-off-by: Hung-Han (Henry) Chen <chenhungh@gmail.com> * Use openai npm Signed-off-by: Hung-Han (Henry) Chen <chenhungh@gmail.com> * Fix generateFromDefaultEndpoint Signed-off-by: Hung-Han (Henry) Chen <chenhungh@gmail.com> * Fix last char become undefined Signed-off-by: Hung-Han (Henry) Chen <chenhungh@gmail.com> * Better support for system prompt Signed-off-by: Hung-Han (Henry) Chen <chenhungh@gmail.com> * Updates Signed-off-by: Hung-Han (Henry) Chen <chenhungh@gmail.com> * Revert Signed-off-by: Hung-Han (Henry) Chen <chenhungh@gmail.com> * Update README Signed-off-by: Hung-Han (Henry) Chen <chenhungh@gmail.com> * Default system prompt Signed-off-by: Hung-Han (Henry) Chen <chenhungh@gmail.com> * remove sk- Signed-off-by: Hung-Han …9db8ced
-
Update websearch prompting & summary prompting (#503)
Unverified signature* Use the task model for websearch query generation * summary with examples in chat format * Update to websearch prompting - Removed web search template - Added examples in `generateQuery` - Passed previous questions when generating the web search query - Added previous questions to the bottom of the context window in `buildPrompt` when using websearch * wait for model if needed * rm console.log * get rid of regex * added a few time related examples * Update src/lib/server/websearch/generateQuery.ts Co-authored-by: Mishig <mishig.davaadorj@coloradocollege.edu> * reformat yesterday * test * fix whitespace * remove extra whitespace for mistral * removed mistral extra whitespace * Update src/lib/server/summarize.ts Co-authored-by: Victor Muštar <victor.mustar@gmail.com> * Update src/lib/server/summarize.ts Co-authored-by: Victor Muštar <victor.mustar@gmail.com> * add an emoji if none is found in the first three characters ----…1e5090f
-
[Websearch] update (#427)
Mishig +2: julien-c, victor Unverified signature* Fix reuqest body * update webSearchQueryPromptTemplate * update generate google query parser * Add today's date to google search query creator * crawl top stories if exts; remove answer_box & knowledgeGraph * Create paragraph chunks from top articles * flattened paragprah chunks * update status texts * add gradio client * call gradio app for RAG * Web scrape only "p, li, span" els * add MAX_N_CHUNKS * gradio result typing * parse only <p> elements * rm dev change * update typing WebSearch * buld RAG prompt * Rm dev change * change websearch context msg from user to assisntat type * use hosted gradio app * fix lint * prompt engineering * more prompt engineering * MAX_N_PAGES_SCRAPE = 10 * better error msg * more prompt engineering * revert websearch prompt to previous * rm `top_stories` from websearch as the results are not good * Stop using gradio client, use regular fetch * chore * Rm websearch…ebac87f
-
Make all prompt templates configurable (#400)
Unverified signature* Make all prompt templates configurable This PR enables all prompts to be configurable using handlebar templates as described in #382. For backward compatibility the old hardcoded templates are reimplemented as the default templates. The old template parameters such as `preprompt`, `userMessageToken`, `userMessageEndToken`, `assistantMessageToken`, `assistantMessageEndToken` are now considered legacy. They still work as they are exposed as variables to the default template. However, new prompt configurations should not use these. And it is recommended that the legacy support is eventually removed. As an example, this is how the default chat prompt template is implemented: ``` {{preprompt}} {{#each messages}} {{#ifUser}}{{@root.userMessageToken}}{{content}}{{@root.userMessageEndToken}}{{/ifUser}} {{#ifAssistant}}{{@root.assistantMessageToken}}{{content}}{{@root.assistantMessageEndToken}}{{/ifAssistant}} {{/each}} {{assistantMessageToken}} ``` In addition, …447c0ca
-
allow different user and assistant end-token (#375)
Unverified signature* allow different user and assistant end-token For models like Llama2, the EndToken is not the same for a userMessage and an assistantMessage. This implements `userMessageEndToken` and `assistantMessageEndToken` which overwrites the messageEndToken behavior. This PR also allows empty strings as userMessageToken and assistantMessageToken and makes this the default. This adds additional flexibility, which is required in the case of Llama2 where the first userMessage is effectively different because of the system message. Note that because `userMessageEndToken` and `assistantMessageToken` are nearly always concatenated, it is almost redundant to have both. The exception is `generateQuery` for websearch which have several consecutive user messages. * Make model branding customizable based on env var (#345) * rm open assistant branding * Update SettingsModal.svelte * make settings work with a dynamic list of models * fixed types --------- Co-authored-…f209301
-
Refactor of websearch (#281)
Unverified signature* broke up websearch into multiple endpoints * Refactored loading to use the load function instead of client side fetching * lint * Chat Logo Home Screen Bookmark icons for iOS (#279) * prettier fix for #279 * fix eslint --------- Co-authored-by: Carolyn Marie <99927612+CarolynM8ri3@users.noreply.github.com>6434339