add: pymupdf/rag

obot-platform · Oct 11, 2024 · ed50874 · ed50874
1 parent 0dddd76
commit ed50874
Show file tree

Hide file tree

Showing 6 changed files with 686 additions and 0 deletions.
diff --git a/LICENSE b/LICENSE
@@ -1,3 +1,5 @@
+DISCLAIMER: The files within the documentloaders/mupdf directory are differently licensed. Please refer to the LICENSE file in that folder for more information.
+
                                  Apache License
                            Version 2.0, January 2004
                         http://www.apache.org/licenses/

diff --git a/documentloaders/mupdf/LICENSE b/documentloaders/mupdf/LICENSE
diff --git a/documentloaders/mupdf/foo.gpt b/documentloaders/mupdf/foo.gpt
@@ -0,0 +1,5 @@
+Tools: ./tool.gpt
+Params: input: Path to input PDF file
+Params: output: Path to output file where the extracted markdown content should be written to. Should end with .md
+
+Call the MuPDF PDF Document Loader tool with the input and output parameters.
diff --git a/documentloaders/mupdf/load.py b/documentloaders/mupdf/load.py
@@ -0,0 +1,7 @@
+import os
+import pymupdf4llm
+import pathlib
+
+md_text = pymupdf4llm.to_markdown(os.getenv("INPUT"))
+
+pathlib.Path(os.getenv("OUTPUT")).write_bytes(md_text.encode())
diff --git a/documentloaders/mupdf/requirements.txt b/documentloaders/mupdf/requirements.txt
@@ -0,0 +1,3 @@
+-i https://pypi.org/simple
+pymupdf==1.24.11; python_version >= '3.8'
+pymupdf4llm==0.0.17
diff --git a/documentloaders/mupdf/tool.gpt b/documentloaders/mupdf/tool.gpt
@@ -0,0 +1,8 @@
+Name: MuPDF PDF Document Loader
+Params: input: Path to input PDF file
+Params: output: Path to output file where the extracted markdown content should be written to. Should end with .md
+metadata: supported-file-type: pdf
+
+#!/usr/bin/env python3 ${GPTSCRIPT_TOOL_DIR}/load.py
+
+