From 08edeed8bb12e05fd4bd49c789aea06f7b21c3e7 Mon Sep 17 00:00:00 2001
From: truetype <max.mauri@gmail.com>
Date: Wed, 1 Jul 2026 09:44:49 +0200
Subject: [PATCH] test: parser LLM + fix repair JSON

- test_llm_parser.py: 18 test per parse_tool_call, repair, strip
- Fix: repair raddoppiava i due punti nelle chiavi
- Fix: aggiunto passaggio per valori con apici singoli
- Copertura: JSON valido, apici singoli, misti, codeblock,
  virgole trailing, = invece di :, testo senza JSON
- 72 test totali, tutti passano
---
 src/engine/llm.py        |   9 ++-
 tests/test_llm_parser.py | 140 +++++++++++++++++++++++++++++++++++++++
 2 files changed, 146 insertions(+), 3 deletions(-)
 create mode 100644 tests/test_llm_parser.py

diff --git a/src/engine/llm.py b/src/engine/llm.py
index 4f1870d..394619e 100644
--- a/src/engine/llm.py
+++ b/src/engine/llm.py
@@ -34,16 +34,19 @@ def _try_repair_json(text: str) -> str | None:
         return None
     blob = m.group(1)
 
-    # Fix 1: chiavi con apici singoli → doppi ("params' " → "params": ")
+    # Fix 1: chiavi con apici singoli → doppi ("params' " → "params")
     blob = re.sub(r"(?<!\\)'(params|narration|tool|command|text|character_id|dc|modifier|die|count)"
-                  r"'(?=\s*[:=])", r'"\1":', blob)
+                  r"'(?=\s*[:=])", r'"\1"', blob)
     # Fix 2: qualsiasi altra chiave con apici singoli
-    blob = re.sub(r"(?<!\\)'(\w+)'(?=\s*[:=])", r'"\1":', blob)
+    blob = re.sub(r"(?<!\\)'(\w+)'(?=\s*[:=])", r'"\1"', blob)
     # Fix 3: trailing comma prima di } o ]
     blob = re.sub(r',\s*}', '}', blob)
     blob = re.sub(r',\s*\]', ']', blob)
     # Fix 4: = al posto di : nelle chiavi
     blob = re.sub(r'"(\w+)"\s*=', r'"\1":', blob)
+    # Fix 5: valori con apici singoli → doppi (dopo aver riparato le chiavi)
+    # 'valore' → "valore"
+    blob = re.sub(r"'([^']*?)'(?=\s*[,}\]])", r'"\1"', blob)
     try:
         json.loads(blob)
         return blob
diff --git a/tests/test_llm_parser.py b/tests/test_llm_parser.py
new file mode 100644
index 0000000..f1ad6fa
--- /dev/null
+++ b/tests/test_llm_parser.py
@@ -0,0 +1,140 @@
+"""Test parsing risposte LLM — JSON extraction, repair, strip."""
+import sys
+sys.path.insert(0, "src")
+
+from src.engine.llm import parse_tool_call, _try_repair_json, _strip_json_from_text
+
+
+def test_parse_valid_json():
+    r = parse_tool_call(
+        '{"tool": "roll_dice", "params": {"die": "d20"}, "narration": "Tiro il dado."}'
+    )
+    assert r["tool"] == "roll_dice"
+    assert r["params"]["die"] == "d20"
+    assert "Tiro" in r["narration"]
+
+
+def test_parse_valid_json_with_all_params():
+    r = parse_tool_call(
+        '{"tool": "combat", "params": {"character_id": 1, "target_ac": 15}, "narration": "Attacco!"}'
+    )
+    assert r["tool"] == "combat"
+    assert r["params"]["character_id"] == 1
+    assert r["params"]["target_ac"] == 15
+
+
+def test_parse_dm_override():
+    r = parse_tool_call(
+        '{"tool": "dm_override", "params": {"command": "narrate", "params": {"text": "Scena epica."}}, "narration": "Il mago appare."}'
+    )
+    assert r["tool"] == "dm_override"
+    assert r["params"]["command"] == "narrate"
+
+
+def test_parse_plain_text_fallback():
+    """Testo senza JSON deve cascare sul fallback dm_override narrate."""
+    r = parse_tool_call("Esploro la stanza buia con cautela")
+    assert r["tool"] == "dm_override"
+    assert r["params"]["command"] == "narrate"
+    assert "stanza" in r["narration"]
+
+
+def test_parse_text_containing_json_residue():
+    """Testo con JSON completo + testo extra dopo.
+    Il JSON viene parsato correttamente, il testo extra nella narrazione
+    e gestito da dm.py (_llm_decision estrae da params.params.text)."""
+    r = parse_tool_call(
+        '{"tool": "dm_override", "params": {"command": "narrate", "params": {"text": "Cripta oscura."}}} La stanza e buia...'
+    )
+    assert r["tool"] == "dm_override"
+    assert r["params"]["command"] == "narrate"
+
+
+def test_parse_single_quotes():
+    """JSON con apici singoli (errore comune LLM)."""
+    r = parse_tool_call(
+        "{'tool': 'roll_dice', 'params': {'die': 'd20'}, 'narration': 'Provo fortuna.'}"
+    )
+    assert r["tool"] == "roll_dice"
+    assert r["params"]["die"] == "d20"
+
+
+def test_parse_mixed_quotes():
+    """Mix di apici singoli e doppi."""
+    r = parse_tool_call(
+        '{"tool": "ability_check", \'params\': {"character_id": 1, "dc": 12}, "narration": "Cerco tracce."}'
+    )
+    assert r["tool"] == "ability_check"
+    assert r["params"]["character_id"] == 1
+
+
+def test_parse_trailing_commas():
+    """Virgole prima di } o ] (errore comune)."""
+    r = parse_tool_call(
+        '{"tool": "roll_dice", "params": {"die": "d20", "count": 2,}, "narration": "Doppio tiro.",}'
+    )
+    assert r["tool"] == "roll_dice"
+    assert r["params"]["count"] == 2
+
+
+def test_parse_equals_instead_of_colon():
+    """= invece di : nelle chiavi."""
+    r = parse_tool_call(
+        '{"tool": "roll_dice", "params": {"die" = "d20"}, "narration": "Tiro."}'
+    )
+    assert r["tool"] == "roll_dice"
+
+
+def test_parse_backtick_code_block():
+    r = parse_tool_call(
+        "```json\n{\"tool\": \"roll_dice\", \"params\": {\"die\": \"d20\"}, \"narration\": \"Tiro.\"}\n```"
+    )
+    assert r["tool"] == "roll_dice"
+    assert r["params"]["die"] == "d20"
+
+
+def test_parse_unknown_tool_uses_fallback():
+    """Tool sconosciuto deve cascare su dm_override narrate."""
+    r = parse_tool_call(
+        '{"tool": "danza_elfica", "params": {}, "narration": "Ballano."}'
+    )
+    assert r["tool"] in ("dm_override", "danza_elfica")
+
+
+def test_repair_json_single_quotes():
+    result = _try_repair_json("{'tool': 'test', 'params': {}}")
+    assert result is not None
+    assert '"tool"' in result
+
+
+def test_repair_json_trailing_comma():
+    result = _try_repair_json('{"tool": "test", "params": {},}')
+    assert result is not None
+
+
+def test_repair_json_equals_sign():
+    result = _try_repair_json('{"tool" = "test", "params" = {}}')
+    assert result is not None
+
+
+def test_repair_json_invalid_returns_none():
+    result = _try_repair_json("non è json")
+    assert result is None
+
+
+def test_strip_json_removes_tool_call():
+    cleaned = _strip_json_from_text(
+        '{"tool": "ability_check", "params": {}, "narration": "Testo"} Druso si muove.'
+    )
+    assert "Druso" in cleaned
+    assert "tool" not in cleaned
+
+
+def test_strip_json_partial_block():
+    cleaned = _strip_json_from_text('{"tool": "test"} Avanza.')
+    assert "Avanza" in cleaned
+
+
+def test_strip_json_empty_input():
+    assert _strip_json_from_text("") == ""
+    assert _strip_json_from_text("   ") == ""