Add a PDF Loader node and explicit PDF support for LoadText (#93)

* pdf loader

* update to use pdfminer

* back to pypdf

Author: kateforsberg13

griptape_nodes_library.json

@@ -2711,6 +2711,17 @@
         "icon": "Combine"
       }
     },
+    {
+      "class_name": "LoadPdf",
+      "file_path": "griptape_nodes_library/files/load_pdf.py",
+      "metadata": {
+        "category": "files",
+        "description": "Load a local PDF file and extract its text content, with outputs for the full joined text, a per-page list, and page count.",
+        "display_name": "Load PDF",
+        "icon": "file-text",
+        "group": "Input/Output"
+      }
+    },
     {
       "class_name": "PathAddExtension",
       "file_path": "griptape_nodes_library/files/path_add_extension.py",
@@ -2840,4 +2851,4 @@
     "workflows/templates/rodin_3d_-_3d_model_from_image.py"
   ],
   "is_default_library": true
-}
+}

griptape_nodes_library/files/load_pdf.py

@@ -0,0 +1,95 @@
+from typing import Any
+
+from griptape.loaders import PdfLoader
+from griptape_nodes.exe_types.core_types import Parameter, ParameterMode
+from griptape_nodes.exe_types.node_types import ControlNode
+from griptape_nodes.traits.file_system_picker import FileSystemPicker
+
+
+class LoadPdf(ControlNode):
+    def __init__(
+        self,
+        name: str,
+        metadata: dict[Any, Any] | None = None,
+    ) -> None:
+        super().__init__(name, metadata)
+
+        self.path = Parameter(
+            name="path",
+            allowed_modes={ParameterMode.INPUT, ParameterMode.PROPERTY},
+            input_types=["str"],
+            type="str",
+            default_value="",
+            tooltip="Path to the local PDF file to load.",
+        )
+        self.path.add_trait(
+            FileSystemPicker(
+                allow_files=True,
+                allow_directories=False,
+                multiple=False,
+                file_types=[".pdf"],
+            )
+        )
+        self.add_parameter(self.path)
+
+        self.add_parameter(
+            Parameter(
+                name="password",
+                allowed_modes={ParameterMode.INPUT, ParameterMode.PROPERTY},
+                input_types=["str"],
+                type="str",
+                default_value="",
+                tooltip="Optional password for password-protected PDFs. Leave blank if not needed.",
+            )
+        )
+
+        self.add_parameter(
+            Parameter(
+                name="text",
+                allowed_modes={ParameterMode.OUTPUT},
+                output_type="str",
+                default_value="",
+                tooltip="The full text content of the PDF, with all pages joined by two newlines.",
+                ui_options={"multiline": True, "placeholder_text": "PDF text will appear here."},
+            )
+        )
+
+        self.add_parameter(
+            Parameter(
+                name="pages",
+                allowed_modes={ParameterMode.OUTPUT},
+                output_type="list",
+                default_value=[],
+                tooltip="A list of strings, one entry per page of the PDF, in order.",
+            )
+        )
+
+        self.add_parameter(
+            Parameter(
+                name="page_count",
+                allowed_modes={ParameterMode.OUTPUT},
+                output_type="int",
+                default_value=0,
+                tooltip="The total number of pages in the PDF.",
+            )
+        )
+
+    def process(self) -> None:
+        path = self.get_parameter_value("path")
+        password = self.get_parameter_value("password") or ""
+
+        loader = PdfLoader()
+        data = loader.fetch(path)
+        pages_artifact = loader.try_parse(data, password=password or None)
+        page_texts = [page.value for page in pages_artifact]
+
+        full_text = "\n\n".join(page_texts)
+        page_count = len(page_texts)
+
+        self.parameter_output_values["text"] = full_text
+        self.parameter_output_values["pages"] = page_texts
+        self.parameter_output_values["page_count"] = page_count
+
+        self.parameter_values["text"] = full_text
+        self.parameter_values["pages"] = page_texts
+        self.parameter_values["page_count"] = page_count

griptape_nodes_library/text/load_text.py

@@ -56,7 +56,8 @@ def process(self) -> None:
         # Load file content based on extension
         ext = os.path.splitext(text_path)[1]  # noqa: PTH122
         if ext.lower() == ".pdf":
-            output_text = PdfLoader().load(text_path)[0].value
+            pages_artifact = PdfLoader().load(text_path)
+            output_text = "\n\n".join(page.value for page in pages_artifact)
         else:
             output_text = File(text_path).read_text()
 

griptape_nodes_library/utils/file_utils.py

@@ -13,6 +13,7 @@
     ".xml",
     ".env",
     ".py",
+    ".pdf",
 }