Test substring fallback on a text-indexed field, document estimator params

A text index cannot serve `substring`, so on a field that has only a text
index the condition runs through the payload fallback; only strict mode may
reject it. Pin that in the OpenAPI suite and reword the strict-mode unit
test comment, which read as if the text index itself blocked the query.

Also spell out what `keys` and `postings` mean in `keys_union_cardinality`.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>

lib/collection/src/problems/unindexed_field.rs

@@ -743,7 +743,9 @@ mod tests {
             infer_index_from_field_condition(&condition),
             vec![FieldIndexType::KeywordMatch],
         );
-        // A text index is not enough: it holds tokens, not raw values.
+        // A text index holds tokens, not raw values, so it cannot serve the
+        // condition and does not satisfy strict mode. Without strict mode the
+        // condition still runs through the payload fallback.
         let text = PayloadFieldSchema::FieldType(PayloadSchemaType::Text);
         assert!(!schema_capabilities(&text).contains(&FieldIndexType::KeywordMatch));
     }

tests/openapi/test_substring_match.py

@@ -155,13 +155,23 @@ def test_substring_match_count():
 
 
 # ---------------------------------------------------------------------------
-# 3. Strict mode: substring filtering requires a keyword index. A text
+# 3. A text index on the field cannot serve substring (it stores tokens, not
-#    index is not enough, since it stores tokens rather than raw values.
+#    raw values), so the condition falls back to the payload scan and still
+#    works. Only strict mode may reject it.
+# ---------------------------------------------------------------------------
+
+def test_substring_match_with_text_index_falls_back():
+    _create_index("description", "text")
+    for substring in SUBSTRING_PROBES:
+        assert _scroll_ids(_substring_filter("description", substring)) == expected_ids(substring), substring
+
+
+# ---------------------------------------------------------------------------
+# 4. Strict mode: substring filtering requires a keyword index. A text
+#    index does not count, since it cannot serve the condition.
 # ---------------------------------------------------------------------------
 
 def test_strict_mode_requires_keyword_index():
-    _create_index("description", "text")
-
     _set_strict_mode({
         "enabled": True,
         "unindexed_filtering_retrieve": False,