feat: 摘要拆分

2023-03-26 22:09:59 +08:00
parent 888642f154
commit 3e4487ad9a
20 changed files with 397 additions and 83 deletions
--- a/src/pages/api/data/postData.ts
+++ b/src/pages/api/data/postData.ts
@@ -2,11 +2,12 @@ import type { NextApiRequest, NextApiResponse } from 'next';
 import { jsonRes } from '@/service/response';
 import { connectToDatabase, Data } from '@/service/mongo';
 import { authToken } from '@/service/utils/tools';
+import type { DataType } from '@/types/data';

 export default async function handler(req: NextApiRequest, res: NextApiResponse) {
  try {
-    let { name } = req.query as { name: string };
-    if (!name) {
+    let { name, type } = req.body as { name: string; type: DataType };
+    if (!name || !type) {
      throw new Error('参数错误');
    }
    await connectToDatabase();
@@ -18,7 +19,8 @@ export default async function handler(req: NextApiRequest, res: NextApiResponse)
    // 生成 data 集合
    const data = await Data.create({
      userId,
-      name
+      name,
+      type
    });

    jsonRes(res, {
--- a/src/pages/api/data/splitData.ts
+++ b/src/pages/api/data/splitData.ts
@@ -1,9 +1,11 @@
 import type { NextApiRequest, NextApiResponse } from 'next';
 import { jsonRes } from '@/service/response';
-import { connectToDatabase, Data, DataItem } from '@/service/mongo';
+import { connectToDatabase, DataItem, Data } from '@/service/mongo';
 import { authToken } from '@/service/utils/tools';
 import { generateQA } from '@/service/events/generateQA';
+import { generateAbstract } from '@/service/events/generateAbstract';

+/* 拆分数据成QA */
 export default async function handler(req: NextApiRequest, res: NextApiResponse) {
  try {
    let { text, dataId } = req.body as { text: string; dataId: string };
@@ -17,14 +19,20 @@ export default async function handler(req: NextApiRequest, res: NextApiResponse)

    const userId = await authToken(authorization);

+    const DataRecord = await Data.findById(dataId);
+
+    if (!DataRecord) {
+      throw new Error('找不到数据集');
+    }
+
    const dataItems: any[] = [];

-    // 格式化文本长度
+    // 每 1000 字符一组
    for (let i = 0; i <= text.length / 1000; i++) {
      dataItems.push({
-        temperature: 0,
        userId,
        dataId,
+        type: DataRecord.type,
        text: text.slice(i * 1000, (i + 1) * 1000),
        status: 1
      });
@@ -33,10 +41,15 @@ export default async function handler(req: NextApiRequest, res: NextApiResponse)
    // 批量插入数据
    await DataItem.insertMany(dataItems);

-    generateQA();
+    try {
+      generateQA();
+      generateAbstract();
+    } catch (error) {
+      error;
+    }

    jsonRes(res, {
-      data: dataItems.length
+      data: ''
    });
  } catch (err) {
    jsonRes(res, {
--- a/src/pages/api/timer/initDataItemTime.ts
+++ b/src/pages/api/timer/initDataItemTime.ts
@@ -13,14 +13,15 @@ export default async function handler(req: NextApiRequest, res: NextApiResponse)
    // await DataItem.updateMany(
    //   {},
    //   {
-    //     times: 2
+    //     type: 'QA'
+    //     // times: 2
    //   }
    // );

    await Data.updateMany(
      {},
      {
-        isDeleted: false
+        type: 'QA'
      }
    );

--- a/src/pages/data/components/CreateDataModal.tsx
+++ b/src/pages/data/components/CreateDataModal.tsx
@@ -8,10 +8,21 @@ import {
  ModalBody,
  ModalCloseButton,
  Button,
-  Input
+  Input,
+  Select,
+  FormControl,
+  FormErrorMessage
 } from '@chakra-ui/react';
 import { postData } from '@/api/data';
 import { useMutation } from '@tanstack/react-query';
+import { useForm, SubmitHandler } from 'react-hook-form';
+import { DataType } from '@/types/data';
+import { DataTypeTextMap } from '@/constants/data';
+
+export interface CreateDataProps {
+  name: string;
+  type: DataType;
+}

 const CreateDataModal = ({
  onClose,
@@ -21,9 +32,20 @@ const CreateDataModal = ({
  onSuccess: () => void;
 }) => {
  const [inputVal, setInputVal] = useState('');
+  const {
+    getValues,
+    register,
+    handleSubmit,
+    formState: { errors }
+  } = useForm<CreateDataProps>({
+    defaultValues: {
+      name: '',
+      type: 'abstract'
+    }
+  });

  const { isLoading, mutate } = useMutation({
-    mutationFn: (name: string) => postData(name),
+    mutationFn: (e: CreateDataProps) => postData(e),
    onSuccess() {
      onSuccess();
      onClose();
@@ -37,23 +59,33 @@ const CreateDataModal = ({
        <ModalHeader>创建数据集</ModalHeader>
        <ModalCloseButton />

-        <ModalBody display={'flex'}>
-          <Input
-            value={inputVal}
-            onChange={(e) => setInputVal(e.target.value)}
-            placeholder={'数据集名称'}
-          ></Input>
+        <ModalBody>
+          <FormControl mb={8} isInvalid={!!errors.name}>
+            <Input
+              placeholder="数据集名称"
+              {...register('name', {
+                required: '数据集名称不能为空'
+              })}
+            />
+            <FormErrorMessage position={'absolute'} fontSize="xs">
+              {!!errors.name && errors.name.message}
+            </FormErrorMessage>
+          </FormControl>
+          <FormControl>
+            <Select placeholder="数据集类型" {...register('type', {})}>
+              {Object.entries(DataTypeTextMap).map(([key, value]) => (
+                <option key={key} value={key}>
+                  {value}
+                </option>
+              ))}
+            </Select>
+          </FormControl>
        </ModalBody>
        <ModalFooter>
          <Button colorScheme={'gray'} onClick={onClose}>
            取消
          </Button>
-          <Button
-            ml={3}
-            isDisabled={inputVal === ''}
-            isLoading={isLoading}
-            onClick={() => mutate(inputVal)}
-          >
+          <Button ml={3} isLoading={isLoading} onClick={handleSubmit(mutate as any)}>
            确认
          </Button>
        </ModalFooter>
--- a/src/pages/data/components/ImportDataModal.tsx
+++ b/src/pages/data/components/ImportDataModal.tsx
@@ -22,6 +22,7 @@ import { useToast } from '@/hooks/useToast';
 import { useLoading } from '@/hooks/useLoading';
 import { formatPrice } from '@/utils/user';
 import { modelList, ChatModelNameEnum } from '@/constants/model';
+import { encode, decode } from 'gpt-token-utils';

 const fileExtension = '.txt,.doc,.docx,.pdf,.md';

@@ -106,6 +107,7 @@ const ImportDataModal = ({
          .join('\n')
          .replace(/\n+/g, '\n');
        setFileText(fileTexts);
+        console.log(encode(fileTexts));
      } catch (error: any) {
        console.log(error);
        toast({
@@ -161,7 +163,9 @@ const ImportDataModal = ({
                  placeholder={'请粘贴或输入需要处理的文本'}
                  onChange={(e) => setTextInput(e.target.value)}
                />
-                <Box mt={2}>一共 {textInput.length} 个字</Box>
+                <Box mt={2}>
+                  一共 {textInput.length} 个字，{encode(textInput).length} 个tokens
+                </Box>
              </>
            )}
            {activeTab === 'doc' && (
@@ -174,12 +178,15 @@ const ImportDataModal = ({
                border={'1px solid '}
                borderColor={'blackAlpha.200'}
                borderRadius={'md'}
+                fontSize={'sm'}
              >
                <Button onClick={onOpen}>选择文件</Button>
                <Box mt={2}>支持 {fileExtension} 文件</Box>
                {fileText && (
                  <>
-                    <Box mt={2}>一共 {fileText.length} 个字</Box>
+                    <Box mt={2}>
+                      一共 {fileText.length} 个字，{encode(fileText).length} 个tokens
+                    </Box>
                    <Box
                      maxH={'300px'}
                      w={'100%'}
--- a/src/pages/data/detail.tsx
+++ b/src/pages/data/detail.tsx
@@ -22,7 +22,7 @@ const DataDetail = ({ dataName, dataId }: { dataName: string; dataId: string })
  return (
    <Card py={4} h={'100%'} display={'flex'} flexDirection={'column'}>
      <Box px={6} fontSize={'xl'} fontWeight={'bold'}>
-        {dataName} 拆分结果
+        {dataName} 结果
      </Box>
      <ScrollData
        flex={'1 0 0'}
@@ -38,8 +38,13 @@ const DataDetail = ({ dataName, dataId }: { dataName: string; dataId: string })
          <Box key={item._id}>
            {item.result.map((result, i) => (
              <Box key={i} mb={3}>
-                <Box fontWeight={'bold'}>Q: {result.q}</Box>
-                <Box>A: {result.a}</Box>
+                {item.type === 'QA' && (
+                  <>
+                    <Box fontWeight={'bold'}>Q: {result.q}</Box>
+                    <Box>A: {result.a}</Box>
+                  </>
+                )}
+                {item.type === 'abstract' && <Box fontSize={'sm'}>{result.abstract}</Box>}
              </Box>
            ))}
          </Box>
--- a/src/pages/data/list.tsx
+++ b/src/pages/data/list.tsx
@@ -28,13 +28,14 @@ import { useRouter } from 'next/router';
 import { useConfirm } from '@/hooks/useConfirm';
 import { useRequest } from '@/hooks/useRequest';
 import { DataItemSchema } from '@/types/mongoSchema';
+import { DataTypeTextMap } from '@/constants/data';
 import { customAlphabet } from 'nanoid';
 const nanoid = customAlphabet('.,', 1);

 const CreateDataModal = dynamic(() => import('./components/CreateDataModal'));
 const ImportDataModal = dynamic(() => import('./components/ImportDataModal'));

-export type ExportDataType = 'jsonl';
+export type ExportDataType = 'jsonl' | 'txt';

 const DataList = () => {
  const router = useRouter();
@@ -84,21 +85,26 @@ const DataList = () => {
      let text = '';
      // 生成 jsonl
      data.forEach((item) => {
-        const result = JSON.stringify({
-          prompt: `${item.q.toLocaleLowerCase()}${nanoid()}</s>`,
-          completion: ` ${item.a}###`
-        });
-        text += `${result}\n`;
+        if (res.type === 'jsonl' && item.q && item.a) {
+          const result = JSON.stringify({
+            prompt: `${item.q.toLocaleLowerCase()}${nanoid()}</s>`,
+            completion: ` ${item.a}###`
+          });
+          text += `${result}\n`;
+        } else if (res.type === 'txt' && item.abstract) {
+          text += `${item.abstract}\n`;
+        }
      });
      // 去掉最后一个 \n
      text = text.substring(0, text.length - 1);
+
      // 导出为文件
      const blob = new Blob([text], { type: 'application/json;charset=utf-8' });

      // 创建下载链接
      const downloadLink = document.createElement('a');
      downloadLink.href = window.URL.createObjectURL(blob);
-      downloadLink.download = 'file.jsonl';
+      downloadLink.download = `data.${res.type}`;

      // 添加链接到页面并触发下载
      document.body.appendChild(downloadLink);
@@ -138,6 +144,7 @@ const DataList = () => {
              <Thead>
                <Tr>
                  <Th>集合名</Th>
+                  <Th>类型</Th>
                  <Th>创建时间</Th>
                  <Th>训练中 / 总数据</Th>
                  <Th></Th>
@@ -158,6 +165,7 @@ const DataList = () => {
                        }}
                      />
                    </Td>
+                    <Td>{DataTypeTextMap[item.type || 'QA']}</Td>
                    <Td>{dayjs(item.createTime).format('YYYY/MM/DD HH:mm')}</Td>
                    <Td>
                      {item.trainingData} / {item.totalData}
@@ -187,9 +195,18 @@ const DataList = () => {
                          导出
                        </MenuButton>
                        <MenuList>
-                          <MenuItem onClick={() => handleExportData({ data: item, type: 'jsonl' })}>
-                            jsonl
-                          </MenuItem>
+                          {item.type === 'QA' && (
+                            <MenuItem
+                              onClick={() => handleExportData({ data: item, type: 'jsonl' })}
+                            >
+                              jsonl
+                            </MenuItem>
+                          )}
+                          {item.type === 'abstract' && (
+                            <MenuItem onClick={() => handleExportData({ data: item, type: 'txt' })}>
+                              txt
+                            </MenuItem>
+                          )}
                        </MenuList>
                      </Menu>

--- a/src/pages/model/components/ModelEditForm.tsx
+++ b/src/pages/model/components/ModelEditForm.tsx
@@ -97,7 +97,7 @@ const ModelEditForm = ({ formHooks }: { formHooks: UseFormReturn<ModelSchema> })
          <Box mb={1}>系统提示词</Box>
          <Textarea
            rows={6}
-            maxLength={500}
+            maxLength={-1}
            {...register('systemPrompt')}
            placeholder={
              '模型默认的 prompt 词，通过调整该内容，可以生成一个限定范围的模型。\n\n注意，改功能会影响对话的整体朝向！'