Catch conditional variable refs and flag duplicate block ids

src/dayamlchecker/yaml_structure.py

@@ -3,6 +3,7 @@
 import argparse
 from dataclasses import dataclass, field
 from pathlib import Path
+from pyexpat import features
 import re
 import sys
 
@@ -1442,6 +1443,95 @@ def _find_unmatched_interview_order_references(
     return unmatched
 
 
+def _extract_mako_guards_by_line(content: str) -> dict[int, list[str]]:
+    """Extract Mako % if conditions active at each line of a content string.
+
+    Returns a dict mapping line number (1-indexed) to list of active guard
+    condition strings at that line.
+    """
+    guards_by_line: dict[int, list[str]] = {}
+    # Stack of (condition_str, start_line) for nested % if blocks
+    guard_stack: list[str] = []
+
+    for i, line in enumerate(content.splitlines(), start=1):
+        stripped = line.strip()
+        if stripped.startswith("% if ") or stripped.startswith("%if "):
+            # Extract condition — everything after "% if " or "%if "
+            cond = re.sub(r"^%\s*if\s+", "", stripped).rstrip(":")
+            guard_stack.append(cond)
+        elif stripped.startswith("% elif ") or stripped.startswith("%elif "):
+            if guard_stack:
+                guard_stack.pop()
+            cond = re.sub(r"^%\s*elif\s+", "", stripped).rstrip(":")
+            guard_stack.append(cond)
+        elif stripped in ("% endif", "%endif", "% endif", "% endfor"):
+            if guard_stack:
+                guard_stack.pop()
+
+        if guard_stack:
+            guards_by_line[i] = list(guard_stack)
+
+    return guards_by_line
+
+
+def _find_unmatched_mako_references(
+    doc: dict[str, Any], conditional_fields: list[dict[str, Any]]
+) -> list[tuple[str, int, str]]:
+    """Check Mako text in question, subquestion, and attachment content blocks for unconditional references to variables that are only conditionally asked"""
+    if not conditional_fields:
+        return []
+
+    # collect (mako_text, source_label) pairs to check
+    text_blocks: list[tuple[str, str]] = []
+
+    question = _get_case_insensitive(doc, "question")
+    if isinstance(question, str):
+        text_blocks.append((question, "question"))
+
+    subquestion = _get_case_insensitive(doc, "subquestion")
+    if isinstance(subquestion, str):
+        text_blocks.append((subquestion, "subquestion"))
+
+    attachment = _get_case_insensitive(doc, "attachment")
+    attachments = _get_case_insensitive(doc, "attachments")
+
+    if isinstance(attachment, dict):
+        content = attachment.get("content")
+        if isinstance(content, str):
+            text_blocks.append((content, "attachment content"))
+    elif isinstance(attachment, list):
+        for item in attachment:
+            if isinstance(item, dict):
+                content = item.get("content")
+                if isinstance(content, str):
+                    text_blocks.append((content, "attachment content"))
+
+    if isinstance(attachments, list):
+        for item in attachments:
+            if isinstance(item, dict):
+                content = item.get("content")
+                if isinstance(content, str):
+                    text_blocks.append((content, "attachment content"))
+
+    if not text_blocks:
+        return []
+
+    unmatched: list[tuple[str, int, str]] = []
+    for text, source in text_blocks:
+        mako_guards_by_line = _extract_mako_guards_by_line(text)
+        for conditional in conditional_fields:
+            field_var = conditional["field_var"]
+            expected_guards = conditional["guards"]
+            ref_lines = _find_variable_reference_lines(text, field_var)
+            for ref_line in ref_lines:
+                active_guards = mako_guards_by_line.get(ref_line, [])
+                if _has_matching_guard(active_guards, expected_guards):
+                    continue
+                unmatched.append((field_var, ref_line, source))
+
+    return unmatched
+
+
 def _max_screen_visibility_nesting_depth(doc: dict[str, Any]) -> int:
     fields = _get_case_insensitive(doc, "fields")
     if not isinstance(fields, list):
@@ -1517,6 +1607,8 @@ def find_errors_from_string(
     ]
     yaml_parser = _make_yaml_parser()
     prior_conditional_fields: list[dict[str, Any]] = []
+    skip_undefined = False
+    seen_ids: dict[str, int] = {}
     line_number = 1
     for source_code in document_match.split(full_content):
         lines_in_code = sum(l == "\n" for l in source_code)
@@ -1640,6 +1732,29 @@ def find_errors_from_string(
                             file_name=input_file,
                         )
                     )
+        features = _get_case_insensitive(doc, "features")
+        if isinstance(features, dict):
+            skip_val = features.get("skip undefined")
+            if skip_val is True or (
+                isinstance(skip_val, str) and skip_val.lower() == "true"
+            ):
+                skip_undefined = True
+
+        block_id = _get_case_insensitive(doc, "id")
+        if isinstance(block_id, str) and block_id.strip():
+            block_id_clean = block_id.strip()
+            block_start = line_number + 1 if line_number > 1 else line_number
+            if block_id_clean in seen_ids:
+                all_errors.append(
+                    YAMLError(
+                        err_str=f'Duplicate block id "{block_id_clean}" - first used at line {seen_ids[block_id_clean]}, reused here. DA will silently use the later block, which is almost certainly a mistake',
+                        line_number=block_start,
+                        file_name=input_file,
+                        experimental=False,
+                    )
+                )
+            else:
+                seen_ids[block_id_clean] = block_start
 
         unmatched_refs = _find_unmatched_interview_order_references(
             doc, prior_conditional_fields
@@ -1656,6 +1771,23 @@ def find_errors_from_string(
                 )
             )
 
+        if not skip_undefined:
+            unmatched_mako_refs = _find_unmatched_mako_references(
+                doc, prior_conditional_fields
+            )
+            for field_var, ref_line, source in unmatched_mako_refs:
+                all_errors.append(
+                    YAMLError(
+                        err_str=(
+                            f'{source} references "{field_var}", but that field is only asked '
+                            f"when certain conditions are met (show if/hide if). If those conditions are "
+                            f"not met, the interview may get stuck"
+                        ),
+                        line_number=doc["__line__"] + line_number + ref_line,
+                        file_name=input_file,
+                    )
+                )
+
         nesting_depth = _max_screen_visibility_nesting_depth(doc)
         if nesting_depth > 2:
             all_errors.append(