Repository navigation
test/playwright jest cicd - #418
Closed
danorlando wants to merge 300 commits into
Closed
danorlando wants to merge 300 commits into
danorlando wants to merge 300 commits into
Conversation
fix(langchain): remove zapierApiKey option from ChatAgent constructor refactor(langchain): update langchain package to v0.0.64 misc(langchain): update test2, test3, and test4 scripts to use --inspect flag
… Google Custom Search API
…tion fix(agent.js): pass progress callback to sendApiMessage function
…nstructor feat(agent.js): add support for environment variable SERPAPI_API_KEY feat(agent.js): add support for environment variable ZAPIER_NLA_API_KEY docs(agent.js): remove commented out code and add comments to clarify code
feat(langchain): add support for GPT-4 model fix(server/index.js): fix uncaughtException handler to ignore 'fetch failed' errors
…e concise message refactor(agent.js): remove unused variable 'errorMessage' refactor(agent.js): change 'result' variable initialization to an empty object instead of null refactor(agent.js): change error message when response generation fails refactor(agent.js): change output message when response generation fails refactor(agent.js): change output message when response generation succeeds
feat(langchain): add test5 script to package.json for running test5.js script
refactor(test3.js, test5.js): remove commented out code and add comments The changes in agent.js are to improve the message that is returned to the user. The word "response" has been changed to "answer" to better reflect the output of the chatbot. The message has also been updated to provide clearer instructions to the user. The changes in test3.js and test5.js are to remove commented out code and add comments to improve readability.
fix: fix typo in BingAI/Settings.jsx feat: add Dockerfile for app containerization docs(google_search.md): add guide for setting up Google Custom Search API key and ID
…mponent docs: update link to system message guidelines in GOOGLE_SEARCH.md feat: add JAILBREAK_INFO.md guide for Bing AI jailbreak mode system message guidelines
style(agent.js): refactor getActions method to accept an input parameter feat(agent.js): add handleChainEnd method to CustomChatAgent class style(customAgent.js): add a new line to the end of the file style(test5.js): comment out unused variable and update input1 variable style(googleSearch.js): change tool name to kebab-case
feat(langchain): add browser tool to ChatAgent in test2.js feat(langchain): add modelOptions to ChatAgent in test2.js feat(langchain): change question in input1 and request article review summary in test5.js
…rentMessageId field feat(askGPTPlugins.js): add default value of null to parentMessageId parameter in ask function
…ptPlugins' feat(endpoints.js): add support for gptPlugins endpoint feat(PresetItem.jsx): add support for gptPlugins endpoint feat(HoverButtons.jsx): add support for gptPlugins endpoint feat(createPayload.ts): add support for gptPlugins endpoint feat(types.ts): add gptPlugins endpoint to EModelEndpoint enum feat(endpoints.js): add gptPlugins endpoint to availableEndpoints selector feat(cleanupPreset.js): add support for gptPlugins endpoint feat(getDefaultConversation.js): add support for gptPlugins endpoint feat(getIcon.jsx): add support for gptPlugins endpoint feat(handleSubmit.js): add support for gptPlugins endpoint
refactor(agent.js): change tool name from 'google-search' to 'google' refactor(agent.js): update description for 'google' tool feat(agent.js): add support for citing sources when using web links in response message fix(agent.js): update error message to not mention error to user feat(agent.js): add unique message ids for user message and response message feat(agent.js): limit number of search results to 5 in 'google' tool refactor(validateTools.js): add console log to show valid tools
…ate model option refactor(askGPTPlugins.js): remove unused imports and variables refactor(askGPTPlugins.js): remove commented code refactor(askGPTPlugins.js): remove unused parameters in ask function feat(ask/index.js): add askGPTPlugins route to router
…ename it to Plugins
feat(askGPTPlugins.js): add support for debug option in endpointOption feat(askGPTPlugins.js): add support for chatGptLabel, promptPrefix, temperature, top_p, presence_penalty, and frequency_penalty in endpointOption feat(askGPTPlugins.js): add support for sending plugin and pluginend events feat(askGPTPlugins.js): add onAgentAction and onChainEnd callbacks to ChatAgent.sendMessage refactor(titleConvo.js): comment out unused imports refactor(validateTools.js): comment out console.log statement refactor(agent.js): change saveMessage to include unfinished property feat(agent.js): add endpoint property to saveConvo call in saveMessageToDatabase feat(askGPTPlugins.js): add validateTools import and use it to validate endpointOption.tools before passing to ChatAgent constructor feat(askGPTPlugins.js
…variable and add support for gptPlugins endpoint fix(MessageHeader.jsx): disable clicking on non-clickable endpoints
…gin name feat(Plugin.jsx): add loading state and display loading spinner feat(Plugin.jsx): add Disclosure component to Plugin component feat(Plugin.jsx): add Disclosure.Panel to Plugin component to display team pricing information feat(Spinner.jsx): add classProp prop to Spinner component to allow for custom styling feat(Landing.jsx): add Plugin component to Landing page for testing testing gpt plugins feat(plugins): Milestone commit - Add formatAction function to format plugin actions. - Add prefix.js file to store the prefix message for ChatAgent. - Update ask function to include plugin object to store plugin data. - Update onAgentAction and onChainEnd functions to format plugin data and send intermediate messages. - Update response object to include plugin data. The `handlers.js` file now includes a `formatAction` function that formats the action object for display in the UI. The `createOnProgress` function now returns a `sendIntermediateMessage` function that sends intermediate messages to the client. feat (client): add support for plugins in messages This commit adds support for plugins in messages. It includes changes to the `handlers.js`, `index.jsx`, `CodeBlock.jsx`, `Message.jsx`, `MessageHeader.jsx`, and `Plugin.jsx` files. The `index.jsx` file now includes a `plugin` property in the `messageHandler` function. The `CodeBlock.jsx` file now includes a `plugin` property that determines the language of the code block. The `Message.jsx` file now includes a `Plugin` component that displays the plugin used in the message. The `MessageHeader.jsx` file now includes a `Plugins` component that displays the enabled plugins. feat(langchain): add OpenAICreateImage tool for generating images based on user prompts fix(langchain): update validateTools to include create-image tool fix(langchain): save plugin data to messageSchema fix(server/routes/askGPTPlugins.js): save userMessage and response to messageSchema feat(langchain): add SelfReflectionTool Add a new tool to the LangChain agent, SelfReflectionTool, which enhances the agent's self-awareness by reflecting on its thoughts before taking action. The tool provides a space for the agent to explore and organize its ideas in response to the user's message. Also, update the prefix message to reflect the changes in the agent's behavior and the way it should engage with the user. The prefix message now emphasizes the use of tools when necessary, and relying on the agent's knowledge for creative requests. It also provides clear instructions on how to use the 'Action' input and how to carry out tasks in the sequence written by the human. Finally, update the OpenAICreateImage tool to return the image URL in markdown format. The tool replaces newlines and spaces in the input text with hyphens to create a valid markdown link. Milestone commit: better error handling with custom output parser, dir and file re-org style(langchain): fix formatting and add comments to prefix.js fix(langchain): remove commented out code in test6.js feat(langchain): reduce maxAttempts from 3 to 2 in CustomChatAgent's buildPromptPrefix method feat(langchain): add null check for result.output in CustomChatAgent's buildPromptPrefix method style(langchain): improve consistency and readability of code This commit improves the consistency and readability of the code in the langchain directory. Specifically, it: - Changes the case of the "Thought" output in the CustomChatAgent class to match the "Thought" output in the SelfReflectionTool class. - Adds a currentDateString property to the CustomChatAgent class to avoid repeating the same code in multiple places. - Updates the prefix in the prefix.js file to match the current objectives of the ChatGPT model. - Changes the description of the OpenAICreateImage tool to request a description of the image to be generated. - Updates the tools used by the ChatAgent in the askGPTPlugins.js file to include the Google and Browser tools instead of the Calculator and Create-Image tools. feat: add wolfram, improve image creation, rename to dall-e
- Update tool-based instructions to use proper Markdown syntax for image URLs - Adjust temperature for modelOptions in CustomChatAgent class - Comment out console.debug statement in CustomChatAgent class - Update prefix in initializeCustomAgent function to use proper line breaks - Update prefix in instructions.js to use proper line breaks and change "user" to "human" - Update input in test6.js to use Ezra Pound instead of Hemingway - Update return statement in OpenAICreateImage class to use "generated-image" as alt-text - Update description in SelfReflectionTool class to provide clearer instructions - Update tools in ask function in askGPTPlugins.js to use only the DALL-E tool and enable debug mode feat(ask): add support for DALL-E tool in formatAction function feat(ask): add support for self-reflection tool in formatAction function feat(Plugin.jsx): add support for self-reflection tool in Plugin component fix(Plugin.jsx): fix Plugin component to not display 'None' when latest is not available
feat(agent.js): pass message parameter to SelfReflectionTool constructor feat(customAgent.js): add longestToolName variable to CustomOutputParser feat(openaiCreateImage.js): replace new lines with spaces in prompt parameter feat(selfReflection.js): add message parameter to SelfReflectionTool constructor feat(selfReflection.js): add placeholder response to selfReflect function
Collaborator
|
Sorry, I just realized I should've merged this first. I can work on the merge conflicts |
This was referenced Jun 11, 2023
Closed
Collaborator
|
Closing after resolving conflicts and continuing the original branch in #478 |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
5 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This PR makes changes to help streamline the development process by providing an automated authentication process for playwright, separating out configuration for local vs. pipeline runs, and setting up CI/CD pipelines for unit and integration tests with github workflows.
Current Issues:
Type of change
How Has This Been Tested?
Test Configuration:
npm run e2e:debug, see that it automatically authenticates.npm run test:clientfrom root folder, see that the unit tests pass as expected.npm run test:apifrom rootnpm run test:cito test the ci scriptnpm run test:cito test the api ci script (note that tests will fail since they get the env vars from github secrets)To get the github action working, the following should be added to the github action secrets (this should be done before merging the PR):
E2E_USER_EMAILandE2E_USER_PASSWORD: this should be a valid user in the db that the MONGO_URI is connecting toCREDS_KEYandCREDS_IVNote: you may opt to remove lines 4 and 5 of the pipelines as it may be unnecessary to also run unit tests when merging in addition to PRs.
Checklist: