文件操作 | BytesRouter

文件操作可以让你处理索引里的单个文件。你能按名字查找文件、用正则在其解析后的内容里 grep、也能读取整个文件的文本 —— 全程不必接触原始文档。

这些端点在 SDK 里位于 client.retrieval.* 之下。

提示

兼容 MCP 的智能体可以通过 LlamaParse MCP 的 findFilesInIndex、grepFileFromIndex 与 readFileFromIndex 三个工具获得同样的能力。

查找文件

按完整文件名或子串匹配查找文件。

Python
TypeScript
Go
Java
CLI

# Exact name match
results = await client.retrieval.find(
    index_id="<your-index-id>",
    file_name="quarterly-report.pdf",
)

# Substring match (case-insensitive)
results = await client.retrieval.find(
    index_id="<your-index-id>",
    file_name_contains="quarterly",
)

for f in results.items:
    print(f"{f.file_id}: {f.file_name}")

// Substring match
const results = await client.retrieval.find({
  index_id: "<your-index-id>",
  file_name_contains: "quarterly",
});

for (const f of results.items) {
  console.log(`${f.file_id}: ${f.file_name}`);
}

// Exact name match
page, err := client.Retrieval.Find(ctx, llamacloud.RetrievalFindParams{
  IndexID:  "<your-index-id>",
  FileName: llamacloud.String("quarterly-report.pdf"),
})
if err != nil {
  log.Fatal(err)
}

// Substring match (case-insensitive)
page, err = client.Retrieval.Find(ctx, llamacloud.RetrievalFindParams{
  IndexID:          "<your-index-id>",
  FileNameContains: llamacloud.String("quarterly"),
})
if err != nil {
  log.Fatal(err)
}

for _, f := range page.Items {
  fmt.Printf("%s: %s\n", f.FileID, f.FileName)
}

import ai.llamaindex.llamacloud.models.retrieval.RetrievalFindPage;
import ai.llamaindex.llamacloud.models.retrieval.RetrievalFindParams;
import ai.llamaindex.llamacloud.models.retrieval.RetrievalFindResponse;

// Substring match (case-insensitive)
RetrievalFindPage results = client.retrieval().find(
    RetrievalFindParams.builder()
        .indexId("<your-index-id>")
        .fileNameContains("quarterly")
        .build());

for (RetrievalFindResponse f : results.items()) {
    System.out.println(f.fileId() + ": " + f.fileName());
}

终端窗口

# Exact name match
llp beta:retrieval find \
  --index-id "<your-index-id>" \
  --file-name "quarterly-report.pdf"

# Substring match (case-insensitive)
llp beta:retrieval find \
  --index-id "<your-index-id>" \
  --file-name-contains "quarterly" \
  | jq -r '"\(.file_id): \(.file_name)"'

在文件内 grep

在文件的解析文本里搜索一个正则模式。适合在文档里定位特定词语、模式或章节。

Python
TypeScript
Go
Java
CLI

found = await client.retrieval.find(
    index_id="<your-index-id>",
    file_name="quarterly-report.pdf",
)
if not found.items:
    raise SystemExit("quarterly-report.pdf is not in the index")
file_id = found.items[0].file_id

results = await client.retrieval.grep(
    index_id="<your-index-id>",
    file_id=file_id,
    pattern="revenue|profit",
    context_chars=100,  # characters of context around each match
)

for match in results.items:
    print(f"Offset {match.start_char}-{match.end_char}: {match.content}")

const found = await client.retrieval.find({
  index_id: "<your-index-id>",
  file_name: "quarterly-report.pdf",
});
if (found.items.length === 0) {
  throw new Error("quarterly-report.pdf is not in the index");
}
const fileId = found.items[0].file_id;

const results = await client.retrieval.grep({
  index_id: "<your-index-id>",
  file_id: fileId,
  pattern: "revenue|profit",
  context_chars: 100,
});

for (const match of results.items) {
  console.log(`Offset ${match.start_char}-${match.end_char}: ${match.content}`);
}

found, err := client.Retrieval.Find(ctx, llamacloud.RetrievalFindParams{
  IndexID:  "<your-index-id>",
  FileName: llamacloud.String("quarterly-report.pdf"),
})
if err != nil {
  log.Fatal(err)
}
if len(found.Items) == 0 {
  log.Fatal("quarterly-report.pdf is not in the index")
}
fileID := found.Items[0].FileID

page, err := client.Retrieval.Grep(ctx, llamacloud.RetrievalGrepParams{
  IndexID:      "<your-index-id>",
  FileID:       fileID,
  Pattern:      "revenue|profit",
  ContextChars: llamacloud.Int(100), // characters of context around each match
})
if err != nil {
  log.Fatal(err)
}

for _, match := range page.Items {
  fmt.Printf("Offset %d-%d: %s\n", match.StartChar, match.EndChar, match.Content)
}

import ai.llamaindex.llamacloud.models.retrieval.RetrievalFindPage;
import ai.llamaindex.llamacloud.models.retrieval.RetrievalFindParams;
import ai.llamaindex.llamacloud.models.retrieval.RetrievalGrepPage;
import ai.llamaindex.llamacloud.models.retrieval.RetrievalGrepParams;
import ai.llamaindex.llamacloud.models.retrieval.RetrievalGrepResponse;

RetrievalFindPage found = client.retrieval().find(
    RetrievalFindParams.builder()
        .indexId("<your-index-id>")
        .fileName("quarterly-report.pdf")
        .build());
if (found.items().isEmpty()) {
    throw new IllegalStateException("quarterly-report.pdf is not in the index");
}
String fileId = found.items().get(0).fileId();

RetrievalGrepPage results = client.retrieval().grep(
    RetrievalGrepParams.builder()
        .indexId("<your-index-id>")
        .fileId(fileId)
        .pattern("revenue|profit")
        .contextChars(100) // characters of context around each match
        .build());

for (RetrievalGrepResponse match : results.items()) {
    System.out.println("Offset " + match.startChar() + "-" + match.endChar() + ": " + match.content());
}

终端窗口

FILE_ID=$(llp beta:retrieval find \
  --index-id "<your-index-id>" \
  --file-name "quarterly-report.pdf" \
  --max-items 1 \
  | jq -r '.file_id // empty')
if [ -z "$FILE_ID" ]; then
  echo "could not find quarterly-report.pdf in the index" >&2
else
  llp beta:retrieval grep \
    --index-id "<your-index-id>" \
    --file-id "$FILE_ID" \
    --pattern "revenue|profit" \
    --context-chars 100 \
    | jq -r '"Offset \(.start_char)-\(.end_char): \(.content)"'
fi

Grep 参数

参数	类型	默认值	说明
`index_id`	`string`	必填	索引 ID。
`file_id`	`string`	必填	要搜索的文件 ID。
`pattern`	`string`	必填	要匹配的正则模式。
`context_chars`	`int`	`null`	每个匹配前后各带多少字符的上下文。

读取文件

读取文件已解析的文本内容。对大文档支持用 offset 与 max_length 分页。

Python
TypeScript
Go
Java
CLI

found = await client.retrieval.find(
    index_id="<your-index-id>",
    file_name="quarterly-report.pdf",
)
if not found.items:
    raise SystemExit("quarterly-report.pdf is not in the index")
file_id = found.items[0].file_id

# Read the full file
result = await client.retrieval.read(
    index_id="<your-index-id>",
    file_id=file_id,
)
print(result.content)

# Read a specific range
result = await client.retrieval.read(
    index_id="<your-index-id>",
    file_id=file_id,
    offset=1000,
    max_length=5000,
)
print(result.content)

const found = await client.retrieval.find({
  index_id: "<your-index-id>",
  file_name: "quarterly-report.pdf",
});
if (found.items.length === 0) {
  throw new Error("quarterly-report.pdf is not in the index");
}
const fileId = found.items[0].file_id;

// Read the full file
const result = await client.retrieval.read({
  index_id: "<your-index-id>",
  file_id: fileId,
});
console.log(result.content);

// Read a specific range
const partial = await client.retrieval.read({
  index_id: "<your-index-id>",
  file_id: fileId,
  offset: 1000,
  max_length: 5000,
});
console.log(partial.content);

found, err := client.Retrieval.Find(ctx, llamacloud.RetrievalFindParams{
  IndexID:  "<your-index-id>",
  FileName: llamacloud.String("quarterly-report.pdf"),
})
if err != nil {
  log.Fatal(err)
}
if len(found.Items) == 0 {
  log.Fatal("quarterly-report.pdf is not in the index")
}
fileID := found.Items[0].FileID

// Read the full file
result, err := client.Retrieval.Read(ctx, llamacloud.RetrievalReadParams{
  IndexID: "<your-index-id>",
  FileID:  fileID,
})
if err != nil {
  log.Fatal(err)
}
fmt.Println(result.Content)

// Read a specific range
partial, err := client.Retrieval.Read(ctx, llamacloud.RetrievalReadParams{
  IndexID:   "<your-index-id>",
  FileID:    fileID,
  Offset:    llamacloud.Int(1000),
  MaxLength: llamacloud.Int(5000),
})
if err != nil {
  log.Fatal(err)
}
fmt.Println(partial.Content)

import ai.llamaindex.llamacloud.models.retrieval.RetrievalFindPage;
import ai.llamaindex.llamacloud.models.retrieval.RetrievalFindParams;
import ai.llamaindex.llamacloud.models.retrieval.RetrievalReadParams;
import ai.llamaindex.llamacloud.models.retrieval.RetrievalReadResponse;

RetrievalFindPage found = client.retrieval().find(
    RetrievalFindParams.builder()
        .indexId("<your-index-id>")
        .fileName("quarterly-report.pdf")
        .build());
if (found.items().isEmpty()) {
    throw new IllegalStateException("quarterly-report.pdf is not in the index");
}
String fileId = found.items().get(0).fileId();

// Read the full file
RetrievalReadResponse result = client.retrieval().read(
    RetrievalReadParams.builder()
        .indexId("<your-index-id>")
        .fileId(fileId)
        .build());
System.out.println(result.content());

// Read a specific range
RetrievalReadResponse partial = client.retrieval().read(
    RetrievalReadParams.builder()
        .indexId("<your-index-id>")
        .fileId(fileId)
        .offset(1000)
        .maxLength(5000)
        .build());
System.out.println(partial.content());

终端窗口

FILE_ID=$(llp beta:retrieval find \
  --index-id "<your-index-id>" \
  --file-name "quarterly-report.pdf" \
  --max-items 1 \
  | jq -r '.file_id // empty')
if [ -z "$FILE_ID" ]; then
  echo "could not find quarterly-report.pdf in the index" >&2
else
  # Read the full file
  llp beta:retrieval read \
    --index-id "<your-index-id>" \
    --file-id "$FILE_ID" \
    | jq -r '.content'

  # Read a specific range
  llp beta:retrieval read \
    --index-id "<your-index-id>" \
    --file-id "$FILE_ID" \
    --offset 1000 \
    --max-length 5000 \
    | jq -r '.content'
fi

读取参数

参数	类型	默认值	说明
`index_id`	`string`	必填	索引 ID。
`file_id`	`string`	必填	要读取的文件 ID。
`offset`	`int`	`0`	起始字符偏移。
`max_length`	`int`	`null`	从偏移起最多返回多少字符。省略则返回整个文件。

索引

数据加载

查询

存储

成本分析

应用

聊天机器人

问答

结构化数据

智能体

Chat Engines

Query Engine

连接器

文档与节点

摄取流水线

节点解析器

大语言模型

提示词

回调

节点后处理器

响应合成器

检索器

路由

结构化输出

现有数据

Presets and Modes

能力

起步

能力

Examples

Azure Deployment

Databases and Queues

LLM Integrations

查找文件

在文件内 grep

Grep 参数

读取文件

读取参数

友情链接