feat: implement multiple file support in Dataset

transaction_parser/parser_benchmark/doctype/parser_benchmark_dataset/parser_benchmark_dataset.json

@@ -7,11 +7,12 @@
  "field_order": [
   "enabled",
   "column_break_title",
-  "section_break_file",
-  "file",
-  "file_type",
-  "column_break_gbap",
+  "files_section",
+  "files",
+  "is_multiple_files",
+  "section_break_txfs",
   "transaction_type",
+  "column_break_xusg",
   "country",
   "section_break_sobg",
   "company",
@@ -59,20 +60,12 @@
    "label": "Enabled"
   },
   {
-   "fieldname": "section_break_file",
-   "fieldtype": "Section Break"
-  },
-  {
-   "fieldname": "file",
-   "fieldtype": "Attach",
-   "label": "File",
-   "print_hide": 1,
-   "reqd": 1
-  },
-  {
-   "fieldname": "file_type",
-   "fieldtype": "Data",
-   "label": "File Type",
+   "allow_on_submit": 1,
+   "default": "0",
+   "fieldname": "is_multiple_files",
+   "fieldtype": "Check",
+   "in_standard_filter": 1,
+   "label": "Multiple Files",
    "read_only": 1
   },
   {
@@ -91,10 +84,6 @@
    "options": "India\nOther",
    "reqd": 1
   },
-  {
-   "fieldname": "column_break_gbap",
-   "fieldtype": "Column Break"
-  },
   {
    "fieldname": "company",
    "fieldtype": "Link",
@@ -115,6 +104,19 @@
    "label": "Party",
    "options": "party_type"
   },
+  {
+   "fieldname": "files_section",
+   "fieldtype": "Section Break",
+   "label": "Files"
+  },
+  {
+   "allow_on_submit": 1,
+   "fieldname": "files",
+   "fieldtype": "Table",
+   "label": "Files",
+   "options": "Parser Benchmark Dataset File",
+   "reqd": 1
+  },
   {
    "fieldname": "processing_section",
    "fieldtype": "Section Break",
@@ -181,7 +183,6 @@
    "label": "Google Gemini Flash-2.5"
   },
   {
-   "depends_on": "eval: doc.file_type === \"PDF\"",
    "fieldname": "pdf_processor_section",
    "fieldtype": "Section Break",
    "label": "PDF Processors"
@@ -273,6 +274,15 @@
   {
    "fieldname": "column_break_aoce",
    "fieldtype": "Column Break"
+  },
+  {
+   "fieldname": "section_break_txfs",
+   "fieldtype": "Section Break",
+   "label": "Transaction Details"
+  },
+  {
+   "fieldname": "column_break_xusg",
+   "fieldtype": "Column Break"
   }
  ],
  "index_web_pages_for_search": 1,
@@ -283,7 +293,7 @@
    "link_fieldname": "dataset"
   }
  ],
- "modified": "2026-03-27 10:32:46.620190",
+ "modified": "2026-03-30 15:01:03.977940",
  "modified_by": "Administrator",
  "module": "Parser Benchmark",
  "name": "Parser Benchmark Dataset",

transaction_parser/parser_benchmark/doctype/parser_benchmark_dataset/parser_benchmark_dataset.py

@@ -34,6 +34,9 @@ class ParserBenchmarkDataset(Document):
     if TYPE_CHECKING:
         from frappe.types import DF
 
+        from transaction_parser.parser_benchmark.doctype.parser_benchmark_dataset_file.parser_benchmark_dataset_file import (
+            ParserBenchmarkDatasetFile,
+        )
         from transaction_parser.parser_benchmark.doctype.parser_benchmark_expected_field.parser_benchmark_expected_field import (
             ParserBenchmarkExpectedField,
         )
@@ -46,10 +49,10 @@ class ParserBenchmarkDataset(Document):
         docling: DF.Check
         enabled: DF.Check
         expected_fields: DF.Table[ParserBenchmarkExpectedField]
-        file: DF.Attach
-        file_type: DF.Data | None
+        files: DF.Table[ParserBenchmarkDatasetFile]
         google_gemini_flash_25: DF.Check
         google_gemini_pro_25: DF.Check
+        is_multiple_files: DF.Check
         naming_series: DF.Literal["PAR-BM-DTS-"]
         ocrmypdf: DF.Check
         openai_gpt_4o: DF.Check
@@ -62,39 +65,27 @@ class ParserBenchmarkDataset(Document):
         transaction_type: DF.Literal["Sales Order", "Expense"]
     # end: auto-generated types
 
-    SUPPORTED_FILE_TYPES = ("PDF", "CSV", "XLSX", "XLS")
-
     def validate(self):
-        self.set_file_type()
-        self.validate_file_type()
+        self.validate_files()
         self.validate_selected_models()
-        self.validate_selected_processors()
         self.validate_expected_fields()
 
-    def set_file_type(self):
-        if self.file_type and not self.has_value_changed("file"):
-            return
+    def before_update_after_submit(self):
+        self.validate_files()
 
-        file_doc = frappe.get_last_doc("File", filters={"file_url": self.file})
-        self.file_type = file_doc.file_type
+    def validate_files(self):
+        """Set file_type for each row and auto-set is_multiple_files."""
+        for row in self.files:
+            if row.file and (not row.file_type or row.has_value_changed("file")):
+                file_doc = frappe.get_last_doc("File", filters={"file_url": row.file})
+                row.file_type = file_doc.file_type
 
-    def validate_file_type(self):
-        if self.file_type not in self.SUPPORTED_FILE_TYPES:
-            frappe.throw(_("Unsupported file type: {0}").format(self.file_type))
+        self.is_multiple_files = len(self.files) > 1
 
     def validate_selected_models(self):
         if not self.get_selected_models():
             frappe.throw(_("Please select at least one AI Model."))
 
-    def validate_selected_processors(self):
-        if self.file_type != "PDF":
-            for field in PDF_PROCESSOR_FIELD_MAP:
-                self.set(field, 0)
-            return
-
-        if not self.get_selected_processors():
-            frappe.throw(_("Please select at least one PDF Processor."))
-
     def validate_expected_fields(self):
         if not self.expected_fields:
             return
@@ -129,6 +120,18 @@ def get_selected_processors(self) -> list[str]:
             label for field, label in PDF_PROCESSOR_FIELD_MAP.items() if self.get(field)
         ]
 
+    def has_pdf_file(self) -> bool:
+        """Check if any file in the child table is a PDF."""
+        return any(row.file_type == "PDF" for row in self.files)
+
+    def get_file_docs(self) -> list:
+        """Return File documents for each row in the files child table."""
+        file_docs = []
+        for row in self.files:
+            file_doc = frappe.get_last_doc("File", filters={"file_url": row.file})
+            file_docs.append(file_doc)
+        return file_docs
+
 
 @frappe.whitelist()
 def run_benchmark(dataset_name: str):
@@ -156,7 +159,7 @@ def create_and_enqueue_benchmark_logs(dataset_name: str) -> list[str]:
     models = dataset.get_selected_models()
     processors = (
         (dataset.get_selected_processors() or [None])
-        if dataset.file_type == "PDF"
+        if dataset.has_pdf_file()
         else [None]
     )
 

transaction_parser/parser_benchmark/doctype/parser_benchmark_dataset_file/__init__.py

@@ -0,0 +1,2 @@
+# Copyright (c) 2026, Resilient Tech and contributors
+# For license information, please see license.txt

transaction_parser/parser_benchmark/doctype/parser_benchmark_dataset_file/parser_benchmark_dataset_file.json

@@ -0,0 +1,43 @@
+{
+ "actions": [],
+ "creation": "2026-03-30 00:00:00",
+ "doctype": "DocType",
+ "engine": "InnoDB",
+ "field_order": [
+  "file",
+  "column_break_yahx",
+  "file_type"
+ ],
+ "fields": [
+  {
+   "fieldname": "file",
+   "fieldtype": "Attach",
+   "in_list_view": 1,
+   "label": "File",
+   "reqd": 1
+  },
+  {
+   "fieldname": "file_type",
+   "fieldtype": "Data",
+   "in_list_view": 1,
+   "label": "File Type",
+   "read_only": 1
+  },
+  {
+   "fieldname": "column_break_yahx",
+   "fieldtype": "Column Break"
+  }
+ ],
+ "istable": 1,
+ "links": [],
+ "modified": "2026-03-30 15:01:27.752102",
+ "modified_by": "Administrator",
+ "module": "Parser Benchmark",
+ "name": "Parser Benchmark Dataset File",
+ "owner": "Administrator",
+ "permissions": [],
+ "row_format": "Dynamic",
+ "sort_field": "modified",
+ "sort_order": "DESC",
+ "states": []
+}

transaction_parser/parser_benchmark/doctype/parser_benchmark_dataset_file/parser_benchmark_dataset_file.py

@@ -0,0 +1,23 @@
+# Copyright (c) 2026, Resilient Tech and contributors
+# For license information, please see license.txt
+
+from frappe.model.document import Document
+
+
+class ParserBenchmarkDatasetFile(Document):
+    # begin: auto-generated types
+    # This code is auto-generated. Do not modify anything in this block.
+
+    from typing import TYPE_CHECKING
+
+    if TYPE_CHECKING:
+        from frappe.types import DF
+
+        file: DF.Attach
+        file_type: DF.Data | None
+        parent: DF.Data
+        parentfield: DF.Data
+        parenttype: DF.Data
+    # end: auto-generated types
+
+    pass

transaction_parser/parser_benchmark/doctype/parser_benchmark_log/parser_benchmark_log.json

@@ -26,7 +26,6 @@
   "column_break_ubhs",
   "file_parsing_tab",
   "pdf_processor",
-  "file_type",
   "column_break_file_metrics",
   "page_limit",
   "section_break_umzr",
@@ -107,7 +106,6 @@
    "read_only": 1
   },
   {
-   "depends_on": "eval: doc.file_type === \"PDF\"",
    "fieldname": "pdf_processor",
    "fieldtype": "Select",
    "label": "PDF Processor",
@@ -121,13 +119,7 @@
    "label": "Total Time (s)",
    "read_only": 1
   },
-  {
-   "fieldname": "file_type",
-   "fieldtype": "Data",
-   "is_virtual": 1,
-   "label": "File Type",
-   "read_only": 1
-  },
+
   {
    "fieldname": "file_parsing_tab",
    "fieldtype": "Tab Break",

transaction_parser/parser_benchmark/doctype/parser_benchmark_log/parser_benchmark_log.py

@@ -97,7 +97,3 @@ def party(self):
     @property
     def page_limit(self):
         return self.get_from_dataset("page_limit") or 0
-
-    @property
-    def file_type(self):
-        return self.get_from_dataset("file_type")

transaction_parser/parser_benchmark/report/transaction_parser_accuracy_analysis/transaction_parser_accuracy_analysis.js

@@ -12,8 +12,6 @@ const AI_MODELS = [
 	"Google Gemini Flash-2.5",
 ];
 
-const FILE_TYPES = ["PDF", "CSV", "XLSX", "XLS"];
-
 const PDF_PROCESSORS = ["OCRMyPDF", "Docling"];
 
 const PARTY_TYPE_MAP = {
@@ -68,12 +66,6 @@ frappe.query_reports["Transaction Parser Accuracy Analysis"] = {
 			fieldtype: "Dynamic Link",
 			options: "party_type",
 		},
-		{
-			fieldname: "file_type",
-			label: __("File Type"),
-			fieldtype: "MultiSelectList",
-			get_data: (txt) => make_options(FILE_TYPES, txt),
-		},
 		{
 			fieldname: "ai_model",
 			label: __("AI Model"),
@@ -92,6 +84,12 @@ frappe.query_reports["Transaction Parser Accuracy Analysis"] = {
 			fieldtype: "Check",
 			default: 0,
 		},
+		{
+			fieldname: "is_multiple_files",
+			label: __("Multiple Files Only"),
+			fieldtype: "Check",
+			default: 0,
+		},
 	],
 };