Skip to content

Agent Zero AI framework

Enhance document_query to support multiple documents

linuztx63b98451821e
diff --git a/prompts/agent.system.tool.document_query.md b/prompts/agent.system.tool.document_query.md
index e6ca44432..9cbc51db3 100644
--- a/prompts/agent.system.tool.document_query.md
+++ b/prompts/agent.system.tool.document_query.md
@@ -1,60 +1,62 @@
-### document_query:
-This tool can be used to read or analyze remote and local documents.
-It can be used to:
- * Get webpage or remote document text content
- * Get local document text content
- * Answer queries about a webpage, remote or local document
-By default, when the "queries" argument is empty, this tool returns the text content of the document retrieved using OCR.
-Additionally, you can pass a list of "queries" - in this case, the tool returns the answers to all the passed queries about the document.
-!!! This is a universal document reader qnd query tool
-!!! Supported document formats: HTML, PDF, Office Documents (word,excel, powerpoint), Textfiles and many more.
+### document_query
+read and analyze remote/local documents get text content or answer questions
+pass a single url/path or a list for multiple documents in "document"
+for web documents use "http://" or "https://"" prefix
+for local files "file://" prefix is optional but full path is required
+if "queries" is empty tool returns document content
+if "queries" is a list of strings tool returns answers
+supports various formats HTML PDF Office Text etc
+usage:
-#### Arguments:
- * "document" (string) : The web address or local path to the document in question. Webdocuments need "http://" or "https://" protocol prefix. For local files the "file:" protocol prefix is optional. Local files MUST be passed with full filesystem path.
- * "queries" (Optional, list[str]) : Optionally, here you can pass one or more queries to be answered (using and/or about) the document
-
-#### Usage example 1:
-##### Request:
-```json
+1 get content
+~~~json
{
"thoughts": [
- "...",
+ "I need to read..."
],
- "headline": "Reading web document content",
+ "headline": "...",
"tool_name": "document_query",
"tool_args": {
- "document": "https://...somexample",
+ "document": "https://.../document"
}
}
-```
-##### Response:
-```plaintext
-... Here is the entire content of the web document requested ...
-```
+~~~
-#### Usage example 2:
-##### Request:
-```json
+2 query document
+~~~json
{
"thoughts": [
- "...",
+ "I need to answer..."
],
- "headline": "Analyzing document to answer specific questions",
+ "headline": "...",
"tool_name": "document_query",
"tool_args": {
- "document": "https://...somexample",
+ "document": "https://.../document",
"queries": [
- "What is the topic?",
- "Who is the audience?"
+ "What is...",
+ "Who is..."
]
}
}
-```
-##### Response:
-```plaintext
-# What is the topic?
-... Description of the document topic ...
+~~~
-# Who is the audience?
-... The intended document audience list with short descriptions ...
-```
+3 query multiple documents
+~~~json
+{
+ "thoughts": [
+ "I need to compare..."
+ ],
+ "headline": "...",
+ "tool_name": "document_query",
+ "tool_args": {
+ "document": [
+ "https://.../document-one",
+ "file:///path/to/document-two"
+ ],
+ "queries": [
+ "Compare the main conclusions...",
+ "What are the key differences..."
+ ]
+ }
+}
+~~~
diff --git a/python/helpers/document_query.py b/python/helpers/document_query.py
index d60a86574..b6bca9b1f 100644
--- a/python/helpers/document_query.py
+++ b/python/helpers/document_query.py
@@ -361,12 +361,16 @@ class DocumentQueryHelper:
self.progress_callback = progress_callback or (lambda x: None)
async def document_qa(
- self, document_uri: str, questions: Sequence[str]
+ self, document_uris: List[str], questions: Sequence[str]
) -> Tuple[bool, str]:
- self.progress_callback(f"Starting Q&A process")
+ self.progress_callback(
+ f"Starting Q&A process for {len(document_uris)} documents"
+ )
- # index document
- _ = await self.document_get_content(document_uri, True)
+ # index documents
+ await asyncio.gather(
+ *[self.document_get_content(uri, True) for uri in document_uris]
+ )
selected_chunks = {}
for question in questions:
self.progress_callback(f"Optimizing query: {question}")
@@ -381,14 +385,18 @@ class DocumentQueryHelper:
)
).strip()
- self.progress_callback(f"Searching document with query: {optimized_query}")
+ self.progress_callback(f"Searching documents with query: {optimized_query}")
+
+ normalized_uris = [self.store.normalize_uri(uri) for uri in document_uris]
+ doc_filter = " or ".join(
+ [f"document_uri == '{uri}'" for uri in normalized_uris]
+ )
- normalized_uri = self.store.normalize_uri(document_uri)
- chunks = await self.store.search_document(
- document_uri=normalized_uri,
+ chunks = await self.store.search_documents(
query=optimized_query,
limit=100,
threshold=DEFAULT_SEARCH_THRESHOLD,
+ filter=doc_filter,
)
self.progress_callback(f"Found {len(chunks)} chunks")
@@ -397,8 +405,8 @@ class DocumentQueryHelper:
selected_chunks[chunk.metadata["id"]] = chunk
if not selected_chunks:
- self.progress_callback(f"No relevant content found in the document")
- content = f"!!! No content found for document: {document_uri} matching queries: {json.dumps(questions)}"
+ self.progress_callback("No relevant content found in the documents")
+ content = f"!!! No content found for documents: {json.dumps(document_uris)} matching queries: {json.dumps(questions)}"
return False, content
self.progress_callback(
diff --git a/python/tools/document_query.py b/python/tools/document_query.py
index 6fca9271a..a05d0f748 100644
--- a/python/tools/document_query.py
+++ b/python/tools/document_query.py
@@ -1,3 +1,5 @@
+import asyncio
+
from python.helpers.tool import Tool, Response
from python.helpers.document_query import DocumentQueryHelper
@@ -5,10 +7,24 @@ from python.helpers.document_query import DocumentQueryHelper
class DocumentQueryTool(Tool):
async def execute(self, **kwargs):
- document_uri = kwargs["document"] or None
- queries = kwargs["queries"] if "queries" in kwargs else [kwargs["query"]] if ("query" in kwargs and kwargs["query"]) else []
- if not isinstance(document_uri, str) or not document_uri:
+ document_uri = kwargs.get("document")
+ document_uris = []
+
+ if isinstance(document_uri, list):
+ document_uris = document_uri
+ elif isinstance(document_uri, str):
+ document_uris = [document_uri]
+
+ if not document_uris:
return Response(message="Error: no document provided", break_loop=False)
+
+ queries = (
+ kwargs["queries"]
+ if "queries" in kwargs
+ else [kwargs["query"]]
+ if ("query" in kwargs and kwargs["query"])
+ else []
+ )
try:
progress = []
@@ -20,9 +36,12 @@ class DocumentQueryTool(Tool):
helper = DocumentQueryHelper(self.agent, progress_callback)
if not queries:
- content = await helper.document_get_content(document_uri)
+ contents = await asyncio.gather(
+ *[helper.document_get_content(uri) for uri in document_uris]
+ )
+ content = "\n\n---\n\n".join(contents)
else:
- _, content = await helper.document_qa(document_uri, queries)
+ _, content = await helper.document_qa(document_uris, queries)
return Response(message=content, break_loop=False)
except Exception as e: # pylint: disable=broad-exception-caught
return Response(message=f"Error processing document: {e}", break_loop=False)