add: example file

libs/megaparse/src/megaparse/examples/parse_file.py

@@ -0,0 +1,14 @@
+from megaparse.formatter.unstructured_formatter.md_formatter import MarkDownFormatter
+from megaparse.megaparse import MegaParse
+from megaparse.parser.unstructured_parser import UnstructuredParser
+
+if __name__ == "__main__":
+    # Parse a file
+    parser = UnstructuredParser()
+    formatter = MarkDownFormatter()
+
+    megaparse = MegaParse(parser=parser, formatters=[formatter])
+
+    file_path = "libs/megaparse/tests/pdf/sample_pdf.pdf"
+    result = megaparse.load(file_path=file_path)
+    print(result)

libs/megaparse/src/megaparse/formatter/structured_output/__init__.py

@@ -0,0 +1,11 @@
+# from typing import List
+
+# from megaparse.formatter.base import BaseFormatter
+# from pydantic import BaseModel
+
+
+# class StructuredFormatter(BaseFormatter):
+#     async def format_string(
+#         self, text: str, file_path: str | None = None, model: BaseModel | None = None
+#     ) -> BaseModel:
+#         raise NotImplementedError()

libs/megaparse/src/megaparse/formatter/table_formatter/llm_table_formatter.py

@@ -3,9 +3,8 @@
 
 from langchain_core.language_models.chat_models import BaseChatModel
 from langchain_core.prompts import ChatPromptTemplate
-from unstructured.documents.elements import Element
-
 from megaparse.formatter.table_formatter import TableFormatter
+from unstructured.documents.elements import Element
 
 
 class SimpleMDTableFormatter(TableFormatter):

libs/megaparse/src/megaparse/formatter/table_formatter/vision_table_formatter.py

@@ -4,12 +4,11 @@
 
 from langchain_core.language_models.chat_models import BaseChatModel
 from langchain_core.messages import HumanMessage
+from megaparse.formatter.table_formatter import TableFormatter
 from pdf2image import convert_from_path
 from PIL import Image
 from unstructured.documents.elements import Element
 
-from megaparse.formatter.table_formatter import TableFormatter
-
 TABLE_OCR_PROMPT = """
 You are tasked with transcribing the content of a table into markdown format. Your goal is to create a well-structured, readable markdown table that accurately represents the original content while adding appropriate formatting.
 Answer uniquely with the parsed table. Do not include the fenced code blocks backticks.

libs/megaparse/src/megaparse/formatter/unstructured_formatter/__init__.py

@@ -1,8 +1,7 @@
 from typing import List
 
-from unstructured.documents.elements import Element
-
 from megaparse.formatter.base import BaseFormatter
+from unstructured.documents.elements import Element
 
 
 class UnstructuredFormatter(BaseFormatter):

libs/megaparse/src/megaparse/megaparse.py

@@ -61,7 +61,7 @@ async def aload(
 
         try:
             parsed_document = await self.parser.convert(file_path=file_path, file=file)
-            # @chloe FIXME: format_checker needs unstructured Elements as input which is to change
+            # @chloe FIXME: format_checker needs unstructured Elements as input which is to change to a megaparse element
             if self.formatters:
                 for formatter in self.formatters:
                     parsed_document = await formatter.format(parsed_document)