diff options
Diffstat (limited to 'src/server/ApiManagers/AssistantManager.ts')
| -rw-r--r-- | src/server/ApiManagers/AssistantManager.ts | 204 |
1 files changed, 132 insertions, 72 deletions
diff --git a/src/server/ApiManagers/AssistantManager.ts b/src/server/ApiManagers/AssistantManager.ts index b42314e41..36468157a 100644 --- a/src/server/ApiManagers/AssistantManager.ts +++ b/src/server/ApiManagers/AssistantManager.ts @@ -7,6 +7,8 @@ import * as uuid from 'uuid'; import { filesDirectory, publicDirectory } from '../SocketData'; import { Method } from '../RouteManager'; import ApiManager, { Registration } from './ApiManager'; +import axios from 'axios'; +import { Chunk } from '../../client/views/nodes/ChatBox/types'; export enum Directory { parsed_files = 'parsed_files', @@ -36,95 +38,153 @@ const readFileAsync = promisify(fs.readFile); export default class AssistantManager extends ApiManager { protected initialize(register: Registration): void { - const openai = new OpenAI({ apiKey: process.env.OPENAI_KEY, dangerouslyAllowBrowser: true }); + const openai = new OpenAI({ + apiKey: process.env._CLIENT_OPENAI_KEY, // Use client key so don't have to set key seperately for client and server. + dangerouslyAllowBrowser: true, + }); register({ method: Method.POST, - subscription: '/uploadPDFToVectorStore', + subscription: '/getWikipediaSummary', secureHandler: async ({ req, res }) => { - const { urls, threadID, assistantID, vector_store_id } = req.body; - - const csvFilesIds: string[] = []; - const otherFileIds: string[] = []; - const allFileIds: string[] = []; - - const fileProcesses = urls.map(async (source: string) => { - const fullPath = path.join(publicDirectory, source); - const fileData = await openai.files.create({ file: createReadStream(fullPath), purpose: 'assistants' }); - allFileIds.push(fileData.id); - if (source.endsWith('.csv')) { - console.log(source); - csvFilesIds.push(fileData.id); - } else { - openai.beta.vectorStores.files.create(vector_store_id, { file_id: fileData.id }); - otherFileIds.push(fileData.id); - } - }); + const { title } = req.body; try { - await Promise.all(fileProcesses).then(() => { - res.send({ vector_store_id: vector_store_id, openai_file_ids: allFileIds }); + const response = await axios.get('https://en.wikipedia.org/w/api.php', { + params: { + action: 'query', + list: 'search', + srsearch: title, + format: 'json', + }, }); - } catch (error) { - res.status(500).send({ error: 'Failed to process files' + error }); + const summary = response.data.query.search[0].snippet; + if (!summary || summary.length === 0 || summary === '' || summary === ' ') { + res.send({ text: 'No article found with that title.' }); + } else { + res.send({ text: summary }); + } + } catch (error: any) { + console.error('Error retrieving article summary from Wikipedia:', error); + res.status(500).send({ error: 'Error retrieving article summary from Wikipedia.', details: error.message }); } }, }); register({ method: Method.POST, - subscription: '/downloadFileFromOpenAI', + subscription: '/createDocument', secureHandler: async ({ req, res }) => { - const { file_id, file_name } = req.body; - //let files_directory: string; - let files_directory = '/files/openAIFiles/'; - switch (file_name.split('.').pop()) { - case 'pdf': - files_directory = '/files/pdfs/'; - break; - case 'csv': - files_directory = '/files/csv/'; - break; - case 'png': - case 'jpg': - case 'jpeg': - files_directory = '/files/images/'; - break; - default: - break; - } + const { file_path } = req.body; + const public_path = path.join(publicDirectory, file_path); + const file_name = path.basename(file_path); + + try { + // Read file data and convert to base64 + const file_data = fs.readFileSync(public_path, { encoding: 'base64' }); + + const response = await axios.post( + 'http://localhost:8080/createDocument', + { + file_data, + file_name, + }, + { + headers: { + 'Content-Type': 'application/json', + }, + } + ); + + const jobId = response.data.job_id; + + // Poll for results + let result; + while (!result) { + await new Promise(resolve => setTimeout(resolve, 5000)); // Wait for 1 second + const resultResponse = await axios.get(`http://localhost:8080/getResult/${jobId}`); + if (resultResponse.status === 200) { + result = resultResponse.data; + } + } + + if (result.chunks && Array.isArray(result.chunks)) { + for (const chunk of result.chunks) { + if (chunk.metadata && (chunk.metadata.type === 'image' || chunk.metadata.type === 'table')) { + let files_directory = '/files/chunk_images/'; + const directory = path.join(publicDirectory, files_directory); + + if (!fs.existsSync(directory)) { + fs.mkdirSync(directory); + } + + const fileName = path.basename(chunk.metadata.file_path); + const filePath = path.join(directory, fileName); + + // Check if base64_data exists + if (chunk.metadata.base64_data) { + // Decode Base64 and save as file + const buffer = Buffer.from(chunk.metadata.base64_data, 'base64'); + await fs.promises.writeFile(filePath, buffer); - const directory = path.join(publicDirectory, files_directory); + // Update the file path in the chunk + chunk.metadata.file_path = path.join(files_directory, fileName); + chunk.metadata.base64_data = undefined; + } else { + console.warn(`No base64_data found for chunk: ${fileName}`); + } + } + } + } else { + console.warn("Result does not contain an iterable 'chunks' property"); + } - if (!fs.existsSync(directory)) { - fs.mkdirSync(directory); + res.send({ document_json: result }); + } catch (error: any) { + console.error('Error communicating with chatbot:', error); + res.status(500).send({ error: 'Failed to communicate with the chatbot', details: error.message }); } - const file = await openai.files.content(file_id); - const new_file_name = `${uuid.v4()}-${file_name}`; - const file_path = path.join(directory, new_file_name); - const file_array_buffer = await file.arrayBuffer(); - const bufferView = new Uint8Array(file_array_buffer); - try { - const written_file = await writeFileAsync(file_path, bufferView); - console.log(written_file); - console.log(file_path); - console.log(file_array_buffer); - console.log(bufferView); - const file_object = new File([bufferView], file_name); - //DashUploadUtils.upload(file_object, 'openAIFiles'); - res.send({ file_path: path.join(files_directory, new_file_name) }); - /* res.send( { - source: "file", - result: { - accessPaths: { - agnostic: {client: path.join('/files/openAIFiles/', `${uuid.v4()}-${file_name}`)} - }, - rawText: "", - duration: 0, - }, - } ); */ - } catch (error) { - res.status(500).send({ error: 'Failed to write file' + error }); + }, + }); + + register({ + method: Method.POST, + subscription: '/formatChunks', + secureHandler: async ({ req, res }) => { + const { relevantChunks } = req.body; + const content: { type: string; text?: string; image_url?: { url: string } }[] = [{ type: 'text', text: '<chunks>' }]; + + for (const chunk of relevantChunks) { + content.push({ + type: 'text', + text: `<chunk chunk_id=${chunk.id} chunk_type=${chunk.metadata.type === 'image' || chunk.metadata.type === 'table' ? 'image' : 'text'}>`, + }); + + if (chunk.metadata.type === 'image' || chunk.metadata.type === 'table') { + try { + const filePath = serverPathToFile(Directory.parsed_files, chunk.metadata.file_path); + const imageBuffer = await readFileAsync(filePath); + const base64Image = imageBuffer.toString('base64'); + if (base64Image) { + content.push({ + type: 'image_url', + image_url: { + url: `data:image/jpeg;base64,${base64Image}`, + }, + }); + } else { + console.log(`Failed to encode image for chunk ${chunk.id}`); + } + } catch (error) { + console.error(`Error reading image file for chunk ${chunk.id}:`, error); + } + } + + content.push({ type: 'text', text: `${chunk.metadata.text}\n</chunk>\n` }); } + + content.push({ type: 'text', text: '</chunks>' }); + + res.send({ formattedChunks: content }); }, }); } |