From debe5953a814f075aec5f878bd338829cd63ca9e Mon Sep 17 00:00:00 2001 From: Sa Zhang <55871322+Nick17t@users.noreply.github.com> Date: Tue, 24 Sep 2024 21:19:49 +0800 Subject: [PATCH] Fix/update jina ai products labels and descriptions (#8730) Co-authored-by: sa zhang --- .../builtin/jina/tools/jina_reader.yaml | 42 +++++++++---------- .../builtin/jina/tools/jina_search.yaml | 31 +++++++------- .../builtin/jina/tools/jina_tokenizer.yaml | 16 +++++-- 3 files changed, 50 insertions(+), 39 deletions(-) diff --git a/api/core/tools/provider/builtin/jina/tools/jina_reader.yaml b/api/core/tools/provider/builtin/jina/tools/jina_reader.yaml index 58ad6d8694..589bc3433d 100644 --- a/api/core/tools/provider/builtin/jina/tools/jina_reader.yaml +++ b/api/core/tools/provider/builtin/jina/tools/jina_reader.yaml @@ -2,14 +2,14 @@ identity: name: jina_reader author: Dify label: - en_US: JinaReader - zh_Hans: JinaReader - pt_BR: JinaReader + en_US: Fetch Single Page + zh_Hans: 获取单页面 + pt_BR: Fetch Single Page description: human: - en_US: Convert any URL to an LLM-friendly input. Experience improved output for your agent and RAG systems at no cost. - zh_Hans: 将任何 URL 转换为 LLM 友好的输入。无需付费即可体验为您的 Agent 和 RAG 系统提供的改进输出。 - pt_BR: Converta qualquer URL em uma entrada amigável ao LLM. Experimente uma saída aprimorada para seus sistemas de agente e RAG sem custo. + en_US: Fetch the target URL (can be a PDF) and convert it into a LLM-friendly markdown. + zh_Hans: 获取目标网址（可以是 PDF），并将其转换为适合大模型处理的 Markdown 格式。 + pt_BR: Busque a URL de destino (que pode ser um PDF) e converta em um Markdown LLM-friendly. llm: A tool for scraping webpages. Input should be a URL. parameters: - name: url @@ -17,13 +17,13 @@ parameters: required: true label: en_US: URL - zh_Hans: 网页链接 + zh_Hans: 网址 pt_BR: URL human_description: - en_US: used for linking to webpages - zh_Hans: 用于链接到网页 - pt_BR: used for linking to webpages - llm_description: url for scraping + en_US: Web link + zh_Hans: 网页链接 + pt_BR: URL da web + llm_description: url para scraping form: llm - name: request_params type: string @@ -31,14 +31,14 @@ parameters: label: en_US: Request params zh_Hans: 请求参数 - pt_BR: Request params + pt_BR: Parâmetros de solicitação human_description: en_US: | request parameters, format: {"key1": "value1", "key2": "value2"} zh_Hans: | 请求参数，格式：{"key1": "value1", "key2": "value2"} pt_BR: | - request parameters, format: {"key1": "value1", "key2": "value2"} + parâmetros de solicitação, formato: {"key1": "value1", "key2": "value2"} llm_description: request parameters form: llm - name: target_selector @@ -51,7 +51,7 @@ parameters: human_description: en_US: css selector for scraping specific elements zh_Hans: css 选择器用于抓取特定元素 - pt_BR: css selector for scraping specific elements + pt_BR: css selector para scraping de elementos específicos llm_description: css selector of the target element to scrape form: form - name: wait_for_selector @@ -64,7 +64,7 @@ parameters: human_description: en_US: css selector for waiting for specific elements zh_Hans: css 选择器用于等待特定元素 - pt_BR: css selector for waiting for specific elements + pt_BR: css selector para aguardar elementos específicos llm_description: css selector of the target element to wait for form: form - name: image_caption @@ -77,8 +77,8 @@ parameters: pt_BR: Legenda da imagem human_description: en_US: "Captions all images at the specified URL, adding 'Image [idx]: [caption]' as an alt tag for those without one. This allows downstream LLMs to interact with the images in activities such as reasoning and summarizing." - zh_Hans: "为指定 URL 上的所有图像添加标题，为没有标题的图像添加“Image [idx]: [caption]”作为 alt 标签。这允许下游 LLM 在推理和总结等活动中与图像进行交互。" - pt_BR: "Captions all images at the specified URL, adding 'Image [idx]: [caption]' as an alt tag for those without one. This allows downstream LLMs to interact with the images in activities such as reasoning and summarizing." + zh_Hans: "为指定 URL 上的所有图像添加标题，为没有标题的图像添加“Image [idx]: [caption]”作为 alt 标签，以支持下游模型的图像交互。" + pt_BR: "Adiciona legendas a todas as imagens na URL especificada, adicionando 'Imagem [idx]: [legenda]' como uma tag alt para aquelas que não têm uma. Isso permite que os modelos LLM inferiores interajam com as imagens em atividades como raciocínio e resumo." llm_description: Captions all images at the specified URL form: form - name: gather_all_links_at_the_end @@ -91,8 +91,8 @@ parameters: pt_BR: Coletar todos os links ao final human_description: en_US: A "Buttons & Links" section will be created at the end. This helps the downstream LLMs or web agents navigating the page or take further actions. - zh_Hans: 最后会创建一个“按钮和链接”部分。这可以帮助下游 LLM 或 Web 代理浏览页面或采取进一步的行动。 - pt_BR: A "Buttons & Links" section will be created at the end. This helps the downstream LLMs or web agents navigating the page or take further actions. + zh_Hans: 末尾将添加“按钮和链接”部分，方便下游模型或网络代理做页面导航或执行进一步操作。 + pt_BR: Um "Botões & Links" section will be created at the end. This helps the downstream LLMs or web agents navigating the page or take further actions. llm_description: Gather all links at the end form: form - name: gather_all_images_at_the_end @@ -105,8 +105,8 @@ parameters: pt_BR: Coletar todas as imagens ao final human_description: en_US: An "Images" section will be created at the end. This gives the downstream LLMs an overview of all visuals on the page, which may improve reasoning. - zh_Hans: 最后会创建一个“图像”部分。这可以让下游的 LLM 概览页面上的所有视觉效果，从而提高推理能力。 - pt_BR: An "Images" section will be created at the end. This gives the downstream LLMs an overview of all visuals on the page, which may improve reasoning. + zh_Hans: 末尾会新增“图片”部分，方便下游模型全面了解页面的视觉内容，提升推理效果。 + pt_BR: Um "Imagens" section will be created at the end. This gives the downstream LLMs an overview of all visuals on the page, which may improve reasoning. llm_description: Gather all images at the end form: form - name: proxy_server diff --git a/api/core/tools/provider/builtin/jina/tools/jina_search.yaml b/api/core/tools/provider/builtin/jina/tools/jina_search.yaml index 2bc70e1be1..e58c639e56 100644 --- a/api/core/tools/provider/builtin/jina/tools/jina_search.yaml +++ b/api/core/tools/provider/builtin/jina/tools/jina_search.yaml @@ -2,13 +2,14 @@ identity: name: jina_search author: Dify label: - en_US: JinaSearch - zh_Hans: JinaSearch - pt_BR: JinaSearch + en_US: Search the web + zh_Hans: 联网搜索 + pt_BR: Search the web description: human: - en_US: Search on the web and get the top 5 results. Useful for grounding using information from the web. - zh_Hans: 在网络上搜索返回前 5 个结果。 + en_US: Search on the public web of a given query and return the top results as LLM-friendly markdown. + zh_Hans: 针对给定的查询在互联网上进行搜索，并以适合大模型处理的 Markdown 格式返回最相关的结果。 + pt_BR: Procurar na web pública de uma consulta fornecida e retornar os melhores resultados como markdown para LLMs. llm: A tool for searching results on the web for grounding. Input should be a simple question. parameters: - name: query @@ -16,11 +17,13 @@ parameters: required: true label: en_US: Question (Query) - zh_Hans: 信息查询 + zh_Hans: 查询 + pt_BR: Pergunta (Consulta) human_description: en_US: used to find information on the web zh_Hans: 在网络上搜索信息 - llm_description: simple question to ask on the web + pt_BR: Usado para encontrar informações na web + llm_description: Pergunta simples para fazer na web form: llm - name: image_caption type: boolean @@ -32,7 +35,7 @@ parameters: pt_BR: Legenda da imagem human_description: en_US: "Captions all images at the specified URL, adding 'Image [idx]: [caption]' as an alt tag for those without one. This allows downstream LLMs to interact with the images in activities such as reasoning and summarizing." - zh_Hans: "为指定 URL 上的所有图像添加标题，为没有标题的图像添加“Image [idx]: [caption]”作为 alt 标签。这允许下游 LLM 在推理和总结等活动中与图像进行交互。" + zh_Hans: "为指定 URL 上的所有图像添加标题，为没有标题的图像添加“Image [idx]: [caption]”作为 alt 标签，以支持下游模型的图像交互。" pt_BR: "Captions all images at the specified URL, adding 'Image [idx]: [caption]' as an alt tag for those without one. This allows downstream LLMs to interact with the images in activities such as reasoning and summarizing." llm_description: Captions all images at the specified URL form: form @@ -46,8 +49,8 @@ parameters: pt_BR: Coletar todos os links ao final human_description: en_US: A "Buttons & Links" section will be created at the end. This helps the downstream LLMs or web agents navigating the page or take further actions. - zh_Hans: 最后会创建一个“按钮和链接”部分。这可以帮助下游 LLM 或 Web 代理浏览页面或采取进一步的行动。 - pt_BR: A "Buttons & Links" section will be created at the end. This helps the downstream LLMs or web agents navigating the page or take further actions. + zh_Hans: 末尾将添加“按钮和链接”部分，汇总页面上的所有链接。方便下游模型或网络代理做页面导航或执行进一步操作。 + pt_BR: Um "Botão & Links" seção será criada no final. Isso ajuda os LLMs ou agentes da web navegando pela página ou executar ações adicionais. llm_description: Gather all links at the end form: form - name: gather_all_images_at_the_end @@ -60,8 +63,8 @@ parameters: pt_BR: Coletar todas as imagens ao final human_description: en_US: An "Images" section will be created at the end. This gives the downstream LLMs an overview of all visuals on the page, which may improve reasoning. - zh_Hans: 最后会创建一个“图像”部分。这可以让下游的 LLM 概览页面上的所有视觉效果，从而提高推理能力。 - pt_BR: An "Images" section will be created at the end. This gives the downstream LLMs an overview of all visuals on the page, which may improve reasoning. + zh_Hans: 末尾会新增“图片”部分，汇总页面上的所有图片。方便下游模型概览页面的视觉内容，提升推理效果。 + pt_BR: Um "Imagens" seção será criada no final. Isso fornece uma visão geral de todas as imagens na página para os LLMs, que pode melhorar a razão. llm_description: Gather all images at the end form: form - name: proxy_server @@ -74,7 +77,7 @@ parameters: human_description: en_US: Use proxy to access URLs zh_Hans: 利用代理访问 URL - pt_BR: Use proxy to access URLs + pt_BR: Usar proxy para acessar URLs llm_description: Use proxy to access URLs form: form - name: no_cache @@ -83,7 +86,7 @@ parameters: default: false label: en_US: Bypass the Cache - zh_Hans: 绕过缓存 + zh_Hans: 是否绕过缓存 pt_BR: Ignorar o cache human_description: en_US: Bypass the Cache diff --git a/api/core/tools/provider/builtin/jina/tools/jina_tokenizer.yaml b/api/core/tools/provider/builtin/jina/tools/jina_tokenizer.yaml index 62a5c7e7ba..74885cdf9a 100644 --- a/api/core/tools/provider/builtin/jina/tools/jina_tokenizer.yaml +++ b/api/core/tools/provider/builtin/jina/tools/jina_tokenizer.yaml @@ -2,11 +2,14 @@ identity: name: jina_tokenizer author: hjlarry label: - en_US: JinaTokenizer + en_US: Segment + zh_Hans: 切分器 + pt_BR: Segment description: human: - en_US: Free API to tokenize text and segment long text into chunks. - zh_Hans: 免费的API可以将文本tokenize，也可以将长文本分割成多个部分。 + en_US: Split long text into chunks and do tokenization. + zh_Hans: 将长文本拆分成小段落，并做分词处理。 + pt_BR: Dividir o texto longo em pedaços e fazer tokenização. llm: Free API to tokenize text and segment long text into chunks. parameters: - name: content @@ -15,6 +18,7 @@ parameters: label: en_US: Content zh_Hans: 内容 + pt_BR: Conteúdo llm_description: the content which need to tokenize or segment form: llm - name: return_tokens @@ -23,18 +27,22 @@ parameters: label: en_US: Return the tokens zh_Hans: 是否返回tokens + pt_BR: Retornar os tokens human_description: en_US: Return the tokens and their corresponding ids in the response. zh_Hans: 返回tokens及其对应的ids。 + pt_BR: Retornar os tokens e seus respectivos ids na resposta. form: form - name: return_chunks type: boolean label: en_US: Return the chunks zh_Hans: 是否分块 + pt_BR: Retornar os chunks human_description: en_US: Chunking the input into semantically meaningful segments while handling a wide variety of text types and edge cases based on common structural cues. - zh_Hans: 将输入分块为具有语义意义的片段，同时根据常见的结构线索处理各种文本类型和边缘情况。 + zh_Hans: 将输入文本分块为语义有意义的片段，同时基于常见的结构线索处理各种文本类型和特殊情况。 + pt_BR: Dividir o texto de entrada em segmentos semanticamente significativos, enquanto lida com uma ampla variedade de tipos de texto e casos de borda com base em pistas estruturais comuns. form: form - name: tokenizer type: select