forked from mckaywrigley/chatbot-ui
-
Notifications
You must be signed in to change notification settings - Fork 0
/
ingest-regs.ts
223 lines (182 loc) · 6.13 KB
/
ingest-regs.ts
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
import { createClient } from "@supabase/supabase-js"
import { Database } from "@/supabase/types"
import OpenAI from "openai"
import { processCSV, processJSON, processMarkdown, processPdf, processTxt } from "./lib/retrieval/processing"
import fs from 'fs';
import path from 'path';
import dotenv from 'dotenv';
dotenv.config({ path: '.env.local' });
const files: any[] = []
process.env.NEXT_PUBLIC_SUPABASE_ANON_KEY = process.env.SUPABASE_SERVICE_ROLE_KEY
const embeddingsProvider: any = "openai"
const supabase = createClient<Database>(
process.env.NEXT_PUBLIC_SUPABASE_URL!,
process.env.SUPABASE_SERVICE_ROLE_KEY!
)
async function loadFiles() {
const stateRegsDir = 'state-regs'
async function readFilesRecursively(directory: string) {
const items = fs.readdirSync(directory, { withFileTypes: true });
console.log(items)
for (const item of items) {
const fullPath = path.join(directory, item.name);
if (item.isDirectory()) {
await readFilesRecursively(fullPath);
} else {
const content = fs.readFileSync(fullPath)
const tokens = content.length
await createFile(item as any, {
name: item.name,
tokens,
file_path: fullPath,
type: "text",
size: fs.statSync(fullPath).size,
description: "",
}, "openai").then(file => {
files.push(file)
})
}
}
}
await readFilesRecursively(stateRegsDir);
}
async function generateAndStoreEmbeddings() {
const openai = new OpenAI({
apiKey: process.env.OPENAI_API_KEY || "",
});
console.log(files)
for (const file of files) {
const fileExtension = file.name.split(".").pop()?.toLowerCase()
const blob = new Blob([fs.readFileSync(file.file_path)], { type: "text/plain" })
let chunks: FileItemChunk[] = []
switch (fileExtension) {
case "csv":
chunks = await processCSV(blob)
break
case "json":
chunks = await processJSON(blob)
break
case "md":
chunks = await processMarkdown(blob)
break
case "pdf":
chunks = await processPdf(blob)
break
case "txt":
chunks = await processTxt(blob)
break
default:
console.log(`Unsupported file type: ${fileExtension}`)
}
let embeddings: any = []
const response = await openai.embeddings.create({
model: "text-embedding-3-small",
input: chunks.map(chunk => chunk.content)
})
embeddings = response.data.map((item: any) => {
return item.embedding
})
const file_items = chunks.map((chunk, index) => ({
file_id: file.id,
user_id: "77908180-3057-4c8b-9dee-559465a903d1",
content: chunk.content,
tokens: chunk.tokens,
openai_embedding:
embeddingsProvider === "openai"
? ((embeddings[index] || null) as any)
: null,
local_embedding:
embeddingsProvider === "local"
? ((embeddings[index] || null) as any)
: null
}))
await supabase.from("file_items").upsert(file_items)
const totalTokens = file_items.reduce((acc, item) => acc + item.tokens, 0)
await supabase
.from("files")
.update({ tokens: totalTokens })
.eq("id", file.id)
}
}
const uploadFile = async (
file: File,
payload: {
path: string
}
) => {
const SIZE_LIMIT = parseInt(
process.env.NEXT_PUBLIC_USER_FILE_SIZE_LIMIT || "10000000"
)
if (file.size > SIZE_LIMIT) {
throw new Error(
`File must be less than ${Math.floor(SIZE_LIMIT / 1000000)}MB`
)
}
const filePath = payload.path.replace(/[^a-z0-9/.]/gi, "_").toLowerCase()
const { error } = await supabase.storage
.from("files")
.upload(filePath, file, {
upsert: true
})
if (error) {
throw new Error(error.message)
}
return filePath
}
const updateFile = async (
fileId: string,
file: TablesUpdate<"files">
) => {
const { data: updatedFile, error } = await supabase
.from("files")
.update(file)
.eq("id", fileId)
.select("*")
.single()
if (error) {
throw new Error(error.message)
}
return updatedFile
}
// For non-docx files
const createFile = async (
file: File,
fileRecord: TablesInsert<"files">,
// workspace_id: string,
embeddingsProvider: "openai" | "local"
) => {
let validFilename = fileRecord.name.replace(/[^a-z0-9.]/gi, "_").toLowerCase()
const extension = file.name.split(".").pop()
const baseName = validFilename.substring(0, validFilename.lastIndexOf("."))
const maxBaseNameLength = 100 - (extension?.length || 0) - 1
if (baseName.length > maxBaseNameLength) {
fileRecord.name = baseName.substring(0, maxBaseNameLength) + "." + extension
} else {
fileRecord.name = baseName + "." + extension
}
fileRecord.user_id = "77908180-3057-4c8b-9dee-559465a903d1"
const { data: createdFile, error } = await supabase
.from("files")
.insert([fileRecord])
.select("*")
.single()
if (error) {
throw new Error(error.message)
}
const filePath = await uploadFile(file, {
path: fileRecord.file_path
})
await updateFile(createdFile.id, {
file_path: filePath
})
return createdFile
}
// After loading files, generate embeddings
loadFiles().then(() => {
console.log("Files loaded, generating embeddings...");
generateAndStoreEmbeddings(files).then(() => {
console.log("Embeddings generated and stored successfully.");
}).catch(error => {
console.error("Failed to generate or store embeddings:", error);
});
});