eval-protocol 0.3.5.dev1__tar.gz → 0.3.5.dev3__tar.gz

{eval_protocol-0.3.5.dev1/eval_protocol.egg-info → eval_protocol-0.3.5.dev3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: eval-protocol
-Version: 0.3.5.dev1
+Version: 0.3.5.dev3
 Summary: The official Python SDK for Eval Protocol (EP.) EP is an open protocol that standardizes how developers author evals for large language model (LLM) applications.
 Author-email: Fireworks AI <info@fireworks.ai>
 License-Expression: MIT

{eval_protocol-0.3.5.dev1 → eval_protocol-0.3.5.dev3}/eval_protocol/_version.py

@@ -8,11 +8,11 @@ import json
 
 version_json = '''
 {
- "date": "2026-01-06T11:14:00-0800",
+ "date": "2026-01-06T11:34:06-0800",
  "dirty": false,
  "error": null,
- "full-revisionid": "38c47583e50493262d74915850e7a4a7e594baf3",
- "version": "0.3.5.dev.1"
+ "full-revisionid": "5a26ef701c03cf6d62d677188dde481a25725a9a",
+ "version": "0.3.5.dev.3"
 }
 '''  # END VERSION_JSON
 

{eval_protocol-0.3.5.dev1 → eval_protocol-0.3.5.dev3}/eval_protocol/cli_commands/create_rft.py

@@ -103,7 +103,7 @@ def _maybe_transform_dataset_jsonl_via_adapter(
         dir=output_dir,
     ) as f:
         for row in eval_rows:
-            f.write(json.dumps(row.model_dump(mode="json"), ensure_ascii=False) + "\n")
+            f.write(json.dumps(row.model_dump(mode="json", exclude_none=True), ensure_ascii=False) + "\n")
         out_path = os.path.abspath(f.name)
     try:
         rel = os.path.relpath(out_path, project_root)
@@ -380,27 +380,15 @@ def _resolve_evaluator(
         selected_tests = _discover_and_select_tests(project_root, non_interactive=non_interactive)
         if not selected_tests:
             return None, None, None, None
+
         if len(selected_tests) != 1:
             if non_interactive and len(selected_tests) > 1:
                 print("Error: Multiple evaluation tests found in --yes (non-interactive) mode.")
                 print("       Please pass --evaluator or --entry to disambiguate.")
-                try:
-                    # Offer candidate evaluator ids for convenience
-                    tests = _discover_tests(project_root)
-                    if tests:
-                        print("       Candidate evaluator ids:")
-                        for t in tests:
-                            func = t.qualname.split(".")[-1]
-                            stem = os.path.splitext(os.path.basename(t.file_path))[0]
-                            cand = _normalize_evaluator_id(f"{stem}-{func}")
-                            print(f"         - {cand}")
-                except Exception:
-                    pass
             else:
                 print("Error: Please select exactly one evaluation test for 'create rft'.")
             return None, None, None, None
 
-        # Derive evaluator_id from user's single selection
         chosen = selected_tests[0]
         func_name = chosen.qualname.split(".")[-1]
         source_file_name = os.path.splitext(os.path.basename(chosen.file_path))[0]

{eval_protocol-0.3.5.dev1 → eval_protocol-0.3.5.dev3}/eval_protocol/cli_commands/utils.py

@@ -278,7 +278,7 @@ def _format_test_choice(test: DiscoveredTest, idx: int) -> str:
 
 
 def _prompt_select_interactive(tests: list[DiscoveredTest]) -> list[DiscoveredTest]:
-    """Interactive selection with arrow keys using questionary."""
+    """Interactive single selection with arrow keys using questionary (Enter selects highlighted)."""
     try:
         import questionary
 
@@ -289,35 +289,32 @@ def _prompt_select_interactive(tests: list[DiscoveredTest]) -> list[DiscoveredTe
             print(f"\nFound 1 test: {_format_test_choice(tests[0], 1)}")
             confirm = questionary.confirm("Select this test?", default=True, style=custom_style).ask()
             if confirm:
-                return tests
+                return [tests[0]]
             else:
                 return []
 
-        # Build checkbox choices
+        # Build single-select choices
         choices = []
         for idx, t in enumerate(tests, 1):
             choice_text = _format_test_choice(t, idx)
-            choices.append(questionary.Choice(title=choice_text, value=idx - 1, checked=False))
+            choices.append(questionary.Choice(title=choice_text, value=idx - 1))
 
         print()
-        selected_indices = questionary.checkbox(
-            "Select evaluation tests to upload:",
+        selected_index = questionary.select(
+            "Select an evaluation test:",
             choices=choices,
             style=custom_style,
             pointer=">",
-            instruction="(↑↓ move, space select, enter confirm)",
+            instruction="(↑↓ move, enter confirm)",
         ).ask()
 
-        if selected_indices is None:  # Ctrl+C
+        if selected_index is None:  # Ctrl+C / Esc
             print("\nUpload cancelled.")
             return []
 
-        if not selected_indices:
-            return []
-
-        selected_tests = [tests[i] for i in selected_indices]
-        print(f"\n✓ Selected {len(selected_tests)} test(s)")
-        return selected_tests
+        chosen = tests[int(selected_index)]
+        print("\n✓ Selected 1 test")
+        return [chosen]
 
     except ImportError:
         # Fallback to simpler implementation
@@ -372,9 +369,10 @@ def _prompt_select_fallback(tests: list[DiscoveredTest]) -> list[DiscoveredTest]
 
 
 def _prompt_select(tests: list[DiscoveredTest], non_interactive: bool) -> list[DiscoveredTest]:
-    """Prompt user to select tests to upload."""
+    """Prompt user to select exactly one test."""
     if non_interactive:
-        return tests
+        # In non-interactive mode, only proceed if unambiguous.
+        return [tests[0]] if len(tests) == 1 else []
 
     return _prompt_select_interactive(tests)
 
@@ -401,7 +399,16 @@ def _discover_and_select_tests(project_root: str, non_interactive: bool) -> Opti
         return None
 
     if not selected_tests:
-        print("No tests selected.")
+        if non_interactive and len(tests) > 1:
+            print("Error: Multiple evaluation tests found in --yes (non-interactive) mode.")
+            print("       Please pass --evaluator or --entry to disambiguate.")
+        else:
+            print("No test selected.")
+        return None
+
+    # Enforce single-select at the helper level.
+    if len(selected_tests) != 1:
+        print("Error: Please select exactly one evaluation test.")
         return None
 
     return selected_tests

{eval_protocol-0.3.5.dev1 → eval_protocol-0.3.5.dev3/eval_protocol.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: eval-protocol
-Version: 0.3.5.dev1
+Version: 0.3.5.dev3
 Summary: The official Python SDK for Eval Protocol (EP.) EP is an open protocol that standardizes how developers author evals for large language model (LLM) applications.
 Author-email: Fireworks AI <info@fireworks.ai>
 License-Expression: MIT