tuieval 0.2.0.dev70__tar.gz → 0.2.0.dev72__tar.gz

{tuieval-0.2.0.dev70 → tuieval-0.2.0.dev72}/CHANGELOG.md

@@ -2,6 +2,7 @@
 
 ## Unreleased
 
+- Retest: the setup checkbox "Rerun existing results" is now "Retest (ask again, even if already answered)", and `tuieval run --retest` is the same (`--force` still works).
 - `tuieval logs` follows the server log of the model running now and moves on with the run; `logs live` shows each answer as it finishes, `--list` every log.
 - Results tables sort by a click on a column name; a second click reverses. Numbers, times and levels sort by value; empty cells stay last.
 - Nvidia and AMD GPUs on Linux: VRAM from `nvidia-smi`, `amd-smi` or `rocm-smi` sizes models (`gpu_memory_gb` overrides it), runs record peak VRAM, and `tuieval doctor` shows the GPU.

{tuieval-0.2.0.dev70 → tuieval-0.2.0.dev72}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.5
 Name: tuieval
-Version: 0.2.0.dev70
+Version: 0.2.0.dev72
 Summary: Evaluate local and frontier LLMs on your own questions: accuracy, speed, tokens and PASS/FAIL verdicts, in the terminal.
 Project-URL: Homepage, https://github.com/ashe-wb/tuieval
 Project-URL: Issues, https://github.com/ashe-wb/tuieval/issues

{tuieval-0.2.0.dev70 → tuieval-0.2.0.dev72}/src/tuieval/_version.py

@@ -18,7 +18,7 @@ version_tuple: tuple[int | str, ...]
 commit_id: str | None
 __commit_id__: str | None
 
-__version__ = version = '0.2.0.dev70'
-__version_tuple__ = version_tuple = (0, 2, 0, 'dev70')
+__version__ = version = '0.2.0.dev72'
+__version_tuple__ = version_tuple = (0, 2, 0, 'dev72')
 
 __commit_id__ = commit_id = None

{tuieval-0.2.0.dev70 → tuieval-0.2.0.dev72}/src/tuieval/run_evals.py

@@ -11,6 +11,7 @@
     tuieval run --only my-model --packs my-pack,other-pack
     tuieval run --tags moe --packs my-pack        # models tagged "moe"
     tuieval run --packs my-pack --tests a,b       # only these tests of a pack (or --tests my-pack:a,other:b)
+    tuieval run --only my-model --packs my-pack --tests a --retest   # ask a question again
     tuieval run --preset nightly                  # a selection saved in the TUI (presets.toml)
     tuieval run --tier smoke --only my-model      # 3 tests per pack, 1 repeat: check paths and flags
     tuieval run --tier smoke --only openrouter:qwen/qwen3-32b   # any OpenRouter model, no models.toml edit
@@ -742,7 +743,8 @@ def main(argv=None):
     p.add_argument("--packs", help="comma-separated packs (default: all)")
     p.add_argument("--tests", help="comma-separated test ids to run instead of the tier's sample: id with one "
                                    "pack in --packs, else pack:id (and --packs defaults to those packs)")
-    p.add_argument("--force", action="store_true", help="rerun packs (or the --tests) that already have current results")
+    p.add_argument("--retest", "--force", action="store_true", dest="force",
+                   help="ask again even if already answered: the packs, or just the --tests (earlier answers go to history/)")
     p.add_argument("--repeat", type=int, help="times each question is asked, any tier (default: 1 for smoke and "
                                                   "screen, the pack's certification count for certify)")
     p.add_argument("--tier", choices=engine.TIERS, default="screen",

{tuieval-0.2.0.dev70 → tuieval-0.2.0.dev72}/src/tuieval/tui.py

@@ -385,6 +385,9 @@ HELP = {
 [b]Models at a time[/b]: how many of the ticked models run side by side (blank: this machine's
 parallel_models in models.toml, default 1). Answers note the models they ran alongside.
 
+[b]Retest[/b]: ask the ticked questions again even if they're already answered (only the tests
+picked with e, if any). Without it, answered questions are skipped. Replaced answers go to history/.
+
 [b]Keys[/b]
   s  start (with a run going: add to it, or queue it)   r  results
   a  add a model (a GGUF, a model id, openrouter:<id>)  m  scan model folders for new GGUFs
@@ -646,7 +649,7 @@ class SetupScreen(Screen):
             yield Input("", id="repeat", type="integer", max_length=2, placeholder="1")
             yield Label("Models at a time")
             yield Input("", id="parallel", type="integer", max_length=1, placeholder="1")
-            yield Checkbox("Rerun existing results", id="force")
+            yield Checkbox("Retest (ask again, even if already answered)", id="force")
         yield Static(id="estimate")
         with Horizontal(id="buttons"):
             yield Button("Start  [s]", id="start", variant="success")
@@ -958,7 +961,7 @@ class SetupScreen(Screen):
         if not run:
             text += "\n[yellow]Nothing selected can run (see skipping).[/yellow]"
         elif rerun:
-            text += "\n[yellow]Everything selected has results; Start runs it again.[/yellow]"
+            text += "\n[yellow]Everything selected has results; Start retests it.[/yellow]"
         resumed = [j for j in run if "done earlier" in j.note]
         if resumed:
             text += f"   [yellow]{len(resumed)} pack(s) continue from earlier results[/yellow]"

{tuieval-0.2.0.dev70 → tuieval-0.2.0.dev72}/tests/test_tuieval.py

@@ -380,6 +380,9 @@ class PickTests(unittest.TestCase):
         hist = os.listdir(os.path.join(self.ws, "results", "m", "history"))
         self.assertTrue(any(f.startswith("apps-") for f in hist), hist)
         self.assertEqual(self.status(), "certified")
+        self.server("oracle")
+        tuieval(self.ws, "run", "--tests", "apps:b", "--retest")                  # --retest = --force
+        self.assertEqual(self.rows(), {"a": True, "b": True, "c": True, "d": True})
 
     def test_cli_errors(self):
         self.server("oracle")