From 8accf4c64c6b5f92feffbf2435a9dd1057239530 Mon Sep 17 00:00:00 2001
From: Eric Hare <ericrhare@gmail.com>
Date: Mon, 19 Aug 2024 15:31:46 -0700
Subject: [PATCH] feat: Add Unstructured Component to Document Loaders (#3308)

* FEAT: Add Unstructured component

* Update Unstructured.py

* [autofix.ci] apply automated fixes

* Switch to FileInput

* Ensure we import the unstructured component

* Small updates to unstructured comp

* Update Unstructured.py

---------

Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>
---
 .../documentloaders/Unstructured.py           | 58 +++++++++++++++++++
 .../components/documentloaders/__init__.py    |  3 +-
 2 files changed, 60 insertions(+), 1 deletion(-)
 create mode 100644 src/backend/base/langflow/components/documentloaders/Unstructured.py

diff --git a/src/backend/base/langflow/components/documentloaders/Unstructured.py b/src/backend/base/langflow/components/documentloaders/Unstructured.py
new file mode 100644
index 000000000..5fdf05c3c
--- /dev/null
+++ b/src/backend/base/langflow/components/documentloaders/Unstructured.py
@@ -0,0 +1,58 @@
+import os
+
+from typing import List
+
+from langflow.custom import Component
+from langflow.inputs import FileInput, SecretStrInput
+from langflow.template import Output
+from langflow.schema import Data
+
+from langchain_community.document_loaders.unstructured import UnstructuredFileLoader
+
+
+class UnstructuredComponent(Component):
+    display_name = "Unstructured"
+    description = "Unstructured data loader"
+    documentation = "https://python.langchain.com/v0.2/docs/integrations/providers/unstructured/"
+    trace_type = "tool"
+    icon = "Unstructured"
+    name = "Unstructured"
+
+    inputs = [
+        FileInput(
+            name="file",
+            display_name="File",
+            required=True,
+            info="The path to the file with which you want to use Unstructured to parse",
+            file_types=["pdf", "docx", "txt"],  # TODO: Support all unstructured file types
+        ),
+        SecretStrInput(
+            name="api_key",
+            display_name="API Key",
+            required=False,
+            info="Unstructured API Key. Create at: https://unstructured.io/ - If not provided, open source library will be used",
+        ),
+    ]
+
+    outputs = [
+        Output(name="data", display_name="Data", method="load_documents"),
+    ]
+
+    def build_unstructured(self) -> UnstructuredFileLoader:
+        os.environ["UNSTRUCTURED_API_KEY"] = self.api_key
+
+        file_paths = [self.file]
+
+        loader = UnstructuredFileLoader(file_paths)
+
+        return loader
+
+    def load_documents(self) -> List[Data]:
+        unstructured = self.build_unstructured()
+
+        documents = unstructured.load()
+        data = [Data.from_document(doc) for doc in documents]  # Using the from_document method of Data
+
+        self.status = data
+
+        return data
diff --git a/src/backend/base/langflow/components/documentloaders/__init__.py b/src/backend/base/langflow/components/documentloaders/__init__.py
index 8f18cd3db..0bed3869d 100644
--- a/src/backend/base/langflow/components/documentloaders/__init__.py
+++ b/src/backend/base/langflow/components/documentloaders/__init__.py
@@ -1,4 +1,5 @@
 from .Confluence import ConfluenceComponent
 from .GitLoader import GitLoaderComponent
+from .Unstructured import UnstructuredComponent
 
-__all__ = ["ConfluenceComponent", "GitLoaderComponent"]
+__all__ = ["ConfluenceComponent", "GitLoaderComponent", "UnstructuredComponent"]