Back to Details

nemo-retriever

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

nemo-retriever

nemo-retriever

The 
retriever
CLI indexes a folder of PDFs into LanceDB (
retriever ingest
) and serves vector search over it (
retriever query
). For any task about searching/answering questions across a folder of PDFs, use this CLI — do not write a custom RAG.
Beyond PDFs and beyond semantic search. 
retriever ingest
also handles images, Office, HTML, TXT, audio, and video — see 
references/setup.md
for the per-format recipe and 
references/install.md
for the install extras (
[multimedia]
, libreoffice, ffmpeg). For non-semantic operations — page filter, verbatim quote with citation, corpus-level aggregate, chart/image caption hits — see 
references/query.md
. Don't fall back to native Read/Grep/Python on non-PDF inputs.
retriever
CLI可将PDF文件夹索引到LanceDB中（
retriever ingest
），并提供基于向量的搜索服务（
retriever query
）。若需对PDF文件夹进行搜索/问答类任务，请使用该CLI——无需自定义RAG。
超越PDF，超越语义搜索。
retriever ingest
还支持图片、Office文档、HTML、TXT、音频和视频文件——各格式处理方法请参考
references/setup.md
，安装扩展组件（
[multimedia]
、libreoffice、ffmpeg）请参考
references/install.md
。对于非语义操作——页面筛选、带引用的原文提取、语料库级聚合、图表/图片标题匹配等，请参考
references/query.md
。处理非PDF输入时，请勿退而使用原生Read/Grep/Python。

Install (if 
retriever
is missing)

安装（若未安装
retriever
）

If 
command -v retriever
returns nothing, follow 
references/install.md
to install the NeMo Retriever Library before proceeding. It prints 
RETRIEVER_VENV=<path>
; substitute that path for 
<RETRIEVER_VENV>
in every example in this skill (setup, query, troubleshooting, and the CLI references).
若执行
command -v retriever
无返回结果，请先按照
references/install.md
安装NeMo Retriever Library。安装完成后会输出
RETRIEVER_VENV=<path>
；在本技能的所有示例（设置、查询、故障排查及CLI参考）中，请将
<RETRIEVER_VENV>
替换为该路径。

Workflow — read the reference for the current phase, then execute

工作流程——先阅读当前阶段的参考文档，再执行操作

Turn typeRead this onceThen execute
Setup turn (first turn — 
./lancedb/nv-ingest.lance
doesn't exist)
references/setup.md
Build the index
Query turn (every subsequent turn — user asks a question)
references/query.md
One 
retriever query
call
Anything errored or returned empty
references/troubleshooting.md
Apply the named recovery; do not improvise
For the full 
retriever ingest
/ 
retriever query
CLI specs, see 
references/cli/ingest.md
and 
references/cli/query.md
. You do not need these for routine turns — 
<RETRIEVER_VENV>/bin/retriever <subcommand> --help
is faster.
Before ingesting a mixed folder, inventory extensions (
find <dir> -name '*.*' | sed 's/.*\.//' | sort -u
) — 
--input-type=auto
silently drops anything outside the supported set. See 
references/troubleshooting.md
"Unsupported file types".
操作类型先阅读以下文档然后执行
设置阶段（首次操作——
./lancedb/nv-ingest.lance
不存在）
references/setup.md
构建索引
查询阶段（后续所有操作——用户提出问题）
references/query.md
调用一次
retriever query
出现错误或返回空结果
references/troubleshooting.md
按照指定方法恢复；请勿自行操作
完整的
retriever ingest
/
retriever query
CLI规范，请参考
references/cli/ingest.md
和
references/cli/query.md
。日常操作无需查阅这些文档——使用
<RETRIEVER_VENV>/bin/retriever <subcommand> --help
更快捷。
在处理混合文件文件夹前，请先统计文件扩展名（执行
find <dir> -name '*.*' | sed 's/.*\.//' | sort -u
）——
--input-type=auto
会自动丢弃不支持的文件类型。详情请参考
references/troubleshooting.md
中的“不支持的文件类型”部分。

Hard limits (apply to every turn)

硬性限制（适用于所有操作阶段）

  • Setup turn: build the index in one shell command (see 
    references/setup.md
    ). STOP after the index lands.
  • Query turn: at most 2 Bash calls — 1 
    retriever query
    , +1 optional targeted text-extract per 
    references/query.md
    . Reply and then STOP.
  • No narration between tool calls. Tokens you emit between calls become input + cached input for every later turn — quadratic cost. Go straight from reading the summary to writing the JSON file.
  • Banned: 
    TodoWrite
    , Glob, Grep, 
    Read
    of whole PDFs, re-running setup, spawning subagents, speculative "confirmation" calls.
Long query turns (5+ tool calls, 1M+ cache-read tokens) cost ~5× a disciplined turn and almost always still produce the wrong answer. Answering partially beats timing out.
  • 设置阶段：通过一条Shell命令构建索引（参考
    references/setup.md
    ）。索引构建完成后立即停止操作。
  • 查询阶段：最多执行2次Bash调用——1次
    retriever query
    ，+1次根据
    references/query.md
    进行的可选定向文本提取。回复后立即停止操作。
  • 工具调用之间无需说明。调用之间输出的内容会成为后续所有操作的输入及缓存输入——会导致成本呈二次方增长。直接从阅读摘要过渡到编写JSON文件即可。
  • 禁止操作：
    TodoWrite
    、Glob、Grep、读取完整PDF、重新执行设置、生成子代理、推测性“确认”调用。
冗长的查询操作（5次以上工具调用、读取100万以上缓存令牌）的成本约为规范操作的5倍，且几乎总会得出错误答案。部分回答强于超时无响应。