Merge pull request #33 from yibeichan/master

improved function based on a larger redcap csv (bridge2ai)

Author: yibeichan

reproschema/redcap2reproschema.py

@@ -6,6 +6,12 @@
 import yaml
 from bs4 import BeautifulSoup
 
+matrix_group_count = {}
+
+
+def clean_header(header):
+    return {k.lstrip("\ufeff"): v for k, v in header.items()}
+
 
 def normalize_condition(condition_str):
     re_parentheses = re.compile(r"\(([0-9]*)\)")
@@ -34,33 +40,60 @@ def process_visibility(data):
     return visibility_obj
 
 
-def parse_field_type_and_value(data, input_type_map):
-    field_type = data.get("Field Type", "")
-
+def parse_field_type_and_value(field, input_type_map):
+    field_type = field.get("Field Type", "")
     input_type = input_type_map.get(field_type, field_type)
 
+    # Initialize the default value type as string
+    value_type = "xsd:string"
+
+    # Map certain field types directly to xsd types
     value_type_map = {
-        "number": "xsd:int",
+        "text": "xsd:string",
         "date_": "xsd:date",
-        "datetime_": "datetime",
-        "time_": "xsd:date",
-        "email": "email",
-        "phone": "phone",
-    }
-    validation_type = data.get("Text Validation Type OR Show Slider Number", "")
-
-    value_type = value_type_map.get(validation_type, "xsd:string")
+        "datetime_": "xsd:dateTime",
+        "time_": "xsd:time",
+        "email": "xsd:string",
+        "phone": "xsd:string",
+    }  # todo: input_type="signature"
+
+    # Get the validation type from the field, if available
+    validation_type = field.get(
+        "Text Validation Type OR Show Slider Number", ""
+    ).strip()
+
+    if validation_type:
+        # Map the validation type to an XSD type if it's in the map
+        value_type = value_type_map.get(validation_type, "xsd:string")
+    elif field_type in ["radio", "dropdown"]:
+        # If there's no validation type, but the field type is radio or dropdown, use xsd:integer
+        value_type = "xsd:integer"
 
     return input_type, value_type
 
 
-def process_choices(choices_str):
+def process_choices(field_type, choices_str):
+    if field_type not in ["radio", "dropdown"]:  # Handle only radio and dropdown types
+        return None
+
     choices = []
     for choice in choices_str.split("|"):
         parts = choice.split(", ")
-        choice_obj = {"schema:value": int(parts[0]), "schema:name": parts[1]}
+        if len(parts) < 2:
+            print(
+                f"Warning: Skipping invalid choice format '{choice}' in a {field_type} field"
+            )
+            continue
+
+        # Try to convert the first part to an integer, if it fails, keep it as a string
+        try:
+            value = int(parts[0])
+        except ValueError:
+            value = parts[0]
+
+        choice_obj = {"name": parts[1], "value": value}
         if len(parts) == 3:
-            # TODO: handle image url
+            # Handle image url
             choice_obj["schema:image"] = f"{parts[2]}.png"
         choices.append(choice_obj)
     return choices
@@ -90,10 +123,12 @@ def parse_html(input_string, default_language="en"):
             text = element.get_text(strip=True)
             if text:
                 result[lang] = text
-        if not result:
+        if not result:  # If no text was extracted
             result[default_language] = soup.get_text(strip=True)
     else:
-        result[default_language] = input_string
+        result[default_language] = soup.get_text(
+            strip=True
+        )  # Use the entire text as default language text
 
     return result
 
@@ -109,9 +144,22 @@ def process_row(
     response_list,
     additional_notes_list,
 ):
+    global matrix_group_count
+    matrix_group_name = field.get("Matrix Group Name", "")
+    if matrix_group_name:
+        matrix_group_count[matrix_group_name] = (
+            matrix_group_count.get(matrix_group_name, 0) + 1
+        )
+        item_id = f"{matrix_group_name}_{matrix_group_count[matrix_group_name]}"
+    else:
+        item_id = field.get("Variable / Field Name", "")
+
     rowData = {
         "@context": schema_context_url,
         "@type": "reproschema:Field",
+        "@id": item_id,
+        "prefLabel": item_id,
+        "description": f"{item_id} of {form_name}",
     }
 
     field_type = field.get("Field Type", "")
@@ -124,8 +172,20 @@ def process_row(
     if value_type:
         rowData["responseOptions"] = {"valueType": value_type}
 
+    if field_type == "yesno":
+        rowData["responseOptions"] = {
+            "valueType": "xsd:boolean",
+            "choices": [{"name": "Yes", "value": 1}, {"name": "No", "value": 0}],
+        }
+
     for key, value in field.items():
-        if schema_map.get(key) == "allow" and value:
+        if (
+            schema_map.get(key) in ["question", "schema:description", "preamble"]
+            and value
+        ):
+            rowData.update({schema_map[key]: parse_html(value)})
+
+        elif schema_map.get(key) == "allow" and value:
             rowData.setdefault("ui", {}).update({schema_map[key]: value.split(", ")})
 
         elif key in ui_list and value:
@@ -139,8 +199,9 @@ def process_row(
             rowData.setdefault("responseOptions", {}).update({schema_map[key]: value})
 
         elif schema_map.get(key) == "choices" and value:
+            # Pass both field_type and value to process_choices
             rowData.setdefault("responseOptions", {}).update(
-                {"choices": process_choices(value)}
+                {"choices": process_choices(field_type, value)}
             )
 
         elif schema_map.get(key) == "scoringLogic" and value:
@@ -159,9 +220,6 @@ def process_row(
                 {"variableName": field["Variable / Field Name"], "isVis": condition}
             )
 
-        elif key in ["question", "schema:description", "preamble"] and value:
-            rowData.update({schema_map[key]: parse_html(value)})
-
         elif key == "Identifier?" and value:
             identifier_val = value.lower() == "y"
             rowData.update(
@@ -190,6 +248,9 @@ def create_form_schema(
     matrix_list,
     scores_list,
 ):
+    # Use a set to track unique items and preserve order
+    unique_order = list(dict.fromkeys(order.get(form_name, [])))
+
     # Construct the JSON-LD structure
     json_ld = {
         "@context": schema_context_url,
@@ -200,7 +261,7 @@ def create_form_schema(
         "schemaVersion": "1.0.0-rc4",
         "version": "0.0.1",
         "ui": {
-            "order": order.get(form_name, []),
+            "order": unique_order,
             "addProperties": bl_list,
             "shuffle": False,
         },
@@ -310,6 +371,7 @@ def process_csv(
     with open(csv_file, mode="r", encoding="utf-8") as csvfile:
         reader = csv.DictReader(csvfile)
         for row in reader:
+            row = clean_header(row)
             form_name = row["Form Name"]
             if form_name not in datas:
                 datas[form_name] = []
@@ -484,7 +546,6 @@ def main():
     parser.add_argument("yaml_file", help="Path to the Reproschema protocol YAML file.")
     args = parser.parse_args()
 
-    # Call the main conversion function
     redcap2reproschema(args.csv_file, args.yaml_file)