chore: merge release/2.157.4 into main

pyproject.toml

@@ -3,7 +3,7 @@ requires = ["setuptools", "wheel"]
 
 [project]
 name = "kili"
-version = "2.156.2"
+version = "2.157.4"
 description = "Python client for Kili Technology labeling tool"
 readme = "README.md"
 authors = [{ name = "Kili Technology", email = "[email protected]" }]

src/kili/__init__.py

@@ -1,3 +1,3 @@
 """Kili Python SDK."""
 
-__version__ = "2.156.2"
+__version__ = "2.157.4"

src/kili/entrypoints/mutations/asset/__init__.py

@@ -40,7 +40,7 @@ class MutationsAsset(BaseOperationEntrypointMixin):
     def append_many_to_dataset(
         self,
         project_id: str,
-        content_array: Optional[Union[List[str], List[dict]]] = None,
+        content_array: Optional[Union[List[str], List[dict], List[List[dict]]]] = None,
         multi_layer_content_array: Optional[List[List[dict]]] = None,
         external_id_array: Optional[List[str]] = None,
         id_array: Optional[List[str]] = None,

src/kili/presentation/client/label.py

@@ -1229,7 +1229,7 @@ def is_rectangle(coco_annotation, coco_image, kili_annotation):
                 self,  # pyright: ignore[reportGeneralTypeIssues]
                 asset_ids=resolved_asset_ids,
                 project_id=ProjectId(project_id),
-                export_type="latest",
+                export_type="normal" if fmt == "llm_v1" else "latest",
                 label_format=fmt,
                 split_option=layout,
                 single_file=single_file,

src/kili/services/asset_import/helpers.py

@@ -2,6 +2,8 @@
 
 import warnings
 
+SEPARATOR = "___"
+
 
 def is_chat_format(data, required_keys):
     """Checks if llm file data is in chat format."""
@@ -43,8 +45,8 @@ def process_json(data):
         if item["content"] is None:
             raise ValueError("Chat item content cannot be null.")
 
-        # Check if the model is null (indicating a prompt)
-        if item["model"] is None:
+        # Check if the prompt comes from a user or a model
+        if item["role"] in ["user", "system"]:
             # If there's an existing prompt being processed, add it to the prompts list
             if current_prompt is not None:
                 transformed_data["prompts"].append(
@@ -68,8 +70,9 @@ def process_json(data):
                     "title": item["role"],
                 }
             )
-            # Collect model for this item
-            models.append(item["model"])
+            # Add model if not None
+            if item["model"] is not None:
+                models.append(item["model"])
 
     if current_prompt is None:
         raise ValueError(
@@ -85,12 +88,12 @@ def process_json(data):
             }
         )
 
-    chat_item_ids = "_".join(item_ids)
+    chat_item_ids = SEPARATOR.join(item_ids)
 
     # Prepare additional_json_metadata
     additional_json_metadata = {
         "chat_id": chat_id,
-        "models": "_".join(models[-2:]),  # Join the last two models
+        "models": SEPARATOR.join(models[-len(completions) :]),  # Join the evaluated models
         "chat_item_ids": chat_item_ids,  # Concatenate all item IDs
         "text": f"Chat_id: {chat_id}\n\nChat_item_ids: {chat_item_ids}",
     }

src/kili/services/export/format/llm/__init__.py

@@ -5,6 +5,7 @@
 from pathlib import Path
 from typing import Dict, List, Optional, Union
 
+from kili.services.asset_import.helpers import SEPARATOR
 from kili.services.export.exceptions import NotCompatibleInputType
 from kili.services.export.format.base import AbstractExporter
 from kili.services.types import Job
@@ -53,23 +54,26 @@ def process(self, assets: List[Dict]) -> List[Dict[str, Union[List[str], str]]]:
     def _process(self, assets: List[Dict]) -> List[Dict[str, Union[List[str], str]]]:
         result = []
         for asset in assets:
-            jobs_config = self.project["jsonInterface"]["jobs"]
-            latest_label = asset["latestLabel"]
             result.append(
                 {
                     "raw_data": _format_raw_data(asset),
                     "status": asset["status"],
                     "external_id": asset["externalId"],
                     "metadata": asset["jsonMetadata"],
                     "labels": [
-                        {
-                            "author": latest_label["author"]["email"],
-                            "created_at": latest_label["createdAt"],
-                            "label_type": latest_label["labelType"],
-                            "label": _format_json_response(
-                                jobs_config, latest_label["jsonResponse"]
-                            ),
-                        }
+                        list(
+                            map(
+                                lambda label: {
+                                    "author": label["author"]["email"],
+                                    "created_at": label["createdAt"],
+                                    "label_type": label["labelType"],
+                                    "label": _format_json_response(
+                                        self.project["jsonInterface"]["jobs"], label["jsonResponse"]
+                                    ),
+                                },
+                                asset["labels"],
+                            )
+                        )
                     ],
                 }
             )
@@ -110,7 +114,7 @@ def _format_raw_data(asset) -> List[Dict]:
         and isinstance(asset["jsonMetadata"]["chat_item_ids"], str)
         and len(asset["jsonMetadata"]["chat_item_ids"]) > 0
     ):
-        chat_items_ids = str.split(asset["jsonMetadata"]["chat_item_ids"], "_")
+        chat_items_ids = str.split(asset["jsonMetadata"]["chat_item_ids"], SEPARATOR)
     else:
         chat_items_ids = []
 
@@ -119,7 +123,7 @@ def _format_raw_data(asset) -> List[Dict]:
         and isinstance(asset["jsonMetadata"]["models"], str)
         and len(asset["jsonMetadata"]["models"]) > 0
     ):
-        models = str.split(asset["jsonMetadata"]["models"], "_")
+        models = str.split(asset["jsonMetadata"]["models"], SEPARATOR)
     else:
         models = []
 

tests/unit/services/asset_import/test_import_llm.py

@@ -45,7 +45,7 @@ def test_upload_from_one_local_file_in_chat_format(self, *_):
             [False],
             [""],
             [
-                '{"chat_id": "6e4094947af4902cd252421aba9a077e8e4402dd", "models": "model-large_model-medium", "chat_item_ids": "0455df65c2d6bb821a9dc9108ac1d79964a0f571_4c28c86c4b22b3397691ce5cf27197fcf7e8fb2d_7326ff17cbfe7e3cb91b008cf0c496fcd17a1074_375b55d44af2c8c801992089c797df8e12605dfb_9231d8819ac96cc8a6c4b7780c301b796c3f8bf2", "text": "Chat_id: 6e4094947af4902cd252421aba9a077e8e4402dd\\n\\nChat_item_ids: 0455df65c2d6bb821a9dc9108ac1d79964a0f571_4c28c86c4b22b3397691ce5cf27197fcf7e8fb2d_7326ff17cbfe7e3cb91b008cf0c496fcd17a1074_375b55d44af2c8c801992089c797df8e12605dfb_9231d8819ac96cc8a6c4b7780c301b796c3f8bf2"}'
+                '{"chat_id": "6e4094947af4902cd252421aba9a077e8e4402dd", "models": "model-large___model-medium", "chat_item_ids": "0455df65c2d6bb821a9dc9108ac1d79964a0f571___4c28c86c4b22b3397691ce5cf27197fcf7e8fb2d___7326ff17cbfe7e3cb91b008cf0c496fcd17a1074___375b55d44af2c8c801992089c797df8e12605dfb___9231d8819ac96cc8a6c4b7780c301b796c3f8bf2", "text": "Chat_id: 6e4094947af4902cd252421aba9a077e8e4402dd\\n\\nChat_item_ids: 0455df65c2d6bb821a9dc9108ac1d79964a0f571___4c28c86c4b22b3397691ce5cf27197fcf7e8fb2d___7326ff17cbfe7e3cb91b008cf0c496fcd17a1074___375b55d44af2c8c801992089c797df8e12605dfb___9231d8819ac96cc8a6c4b7780c301b796c3f8bf2"}'
             ],
         )
         self.kili.graphql_client.execute.assert_called_with(*expected_parameters)
@@ -69,7 +69,7 @@ def test_upload_from_one_local_file_in_chat_format_with_given_json_metadata(self
             [False],
             [""],
             [
-                '{"customKey": "customValue", "chat_id": "6e4094947af4902cd252421aba9a077e8e4402dd", "models": "model-large_model-medium", "chat_item_ids": "0455df65c2d6bb821a9dc9108ac1d79964a0f571_4c28c86c4b22b3397691ce5cf27197fcf7e8fb2d_7326ff17cbfe7e3cb91b008cf0c496fcd17a1074_375b55d44af2c8c801992089c797df8e12605dfb_9231d8819ac96cc8a6c4b7780c301b796c3f8bf2", "text": "Chat_id: 6e4094947af4902cd252421aba9a077e8e4402dd\\n\\nChat_item_ids: 0455df65c2d6bb821a9dc9108ac1d79964a0f571_4c28c86c4b22b3397691ce5cf27197fcf7e8fb2d_7326ff17cbfe7e3cb91b008cf0c496fcd17a1074_375b55d44af2c8c801992089c797df8e12605dfb_9231d8819ac96cc8a6c4b7780c301b796c3f8bf2"}'
+                '{"customKey": "customValue", "chat_id": "6e4094947af4902cd252421aba9a077e8e4402dd", "models": "model-large___model-medium", "chat_item_ids": "0455df65c2d6bb821a9dc9108ac1d79964a0f571___4c28c86c4b22b3397691ce5cf27197fcf7e8fb2d___7326ff17cbfe7e3cb91b008cf0c496fcd17a1074___375b55d44af2c8c801992089c797df8e12605dfb___9231d8819ac96cc8a6c4b7780c301b796c3f8bf2", "text": "Chat_id: 6e4094947af4902cd252421aba9a077e8e4402dd\\n\\nChat_item_ids: 0455df65c2d6bb821a9dc9108ac1d79964a0f571___4c28c86c4b22b3397691ce5cf27197fcf7e8fb2d___7326ff17cbfe7e3cb91b008cf0c496fcd17a1074___375b55d44af2c8c801992089c797df8e12605dfb___9231d8819ac96cc8a6c4b7780c301b796c3f8bf2"}'
             ],
         )
         self.kili.graphql_client.execute.assert_called_with(*expected_parameters)
@@ -155,7 +155,7 @@ def test_upload_from_dict_in_chat_format(self, *_):
             [False],
             [""],
             [
-                '{"chat_id": "6e4094947af4902cd252421aba9a077e8e4402dd", "models": "model-large_model-medium", "chat_item_ids": "0455df65c2d6bb821a9dc9108ac1d79964a0f571_4c28c86c4b22b3397691ce5cf27197fcf7e8fb2d_7326ff17cbfe7e3cb91b008cf0c496fcd17a1074_375b55d44af2c8c801992089c797df8e12605dfb_9231d8819ac96cc8a6c4b7780c301b796c3f8bf2", "text": "Chat_id: 6e4094947af4902cd252421aba9a077e8e4402dd\\n\\nChat_item_ids: 0455df65c2d6bb821a9dc9108ac1d79964a0f571_4c28c86c4b22b3397691ce5cf27197fcf7e8fb2d_7326ff17cbfe7e3cb91b008cf0c496fcd17a1074_375b55d44af2c8c801992089c797df8e12605dfb_9231d8819ac96cc8a6c4b7780c301b796c3f8bf2"}'
+                '{"chat_id": "6e4094947af4902cd252421aba9a077e8e4402dd", "models": "model-large___model-medium", "chat_item_ids": "0455df65c2d6bb821a9dc9108ac1d79964a0f571___4c28c86c4b22b3397691ce5cf27197fcf7e8fb2d___7326ff17cbfe7e3cb91b008cf0c496fcd17a1074___375b55d44af2c8c801992089c797df8e12605dfb___9231d8819ac96cc8a6c4b7780c301b796c3f8bf2", "text": "Chat_id: 6e4094947af4902cd252421aba9a077e8e4402dd\\n\\nChat_item_ids: 0455df65c2d6bb821a9dc9108ac1d79964a0f571___4c28c86c4b22b3397691ce5cf27197fcf7e8fb2d___7326ff17cbfe7e3cb91b008cf0c496fcd17a1074___375b55d44af2c8c801992089c797df8e12605dfb___9231d8819ac96cc8a6c4b7780c301b796c3f8bf2"}'
             ],
         )
         self.kili.graphql_client.execute.assert_called_with(*expected_parameters)
@@ -213,7 +213,7 @@ def test_upload_from_dict_in_chat_format_with_json_metadata(self, *_):
             [False],
             [""],
             [
-                '{"customKey": "customValue", "chat_id": "6e4094947af4902cd252421aba9a077e8e4402dd", "models": "model-large_model-medium", "chat_item_ids": "0455df65c2d6bb821a9dc9108ac1d79964a0f571_4c28c86c4b22b3397691ce5cf27197fcf7e8fb2d_7326ff17cbfe7e3cb91b008cf0c496fcd17a1074_375b55d44af2c8c801992089c797df8e12605dfb_9231d8819ac96cc8a6c4b7780c301b796c3f8bf2", "text": "Chat_id: 6e4094947af4902cd252421aba9a077e8e4402dd\\n\\nChat_item_ids: 0455df65c2d6bb821a9dc9108ac1d79964a0f571_4c28c86c4b22b3397691ce5cf27197fcf7e8fb2d_7326ff17cbfe7e3cb91b008cf0c496fcd17a1074_375b55d44af2c8c801992089c797df8e12605dfb_9231d8819ac96cc8a6c4b7780c301b796c3f8bf2"}'
+                '{"customKey": "customValue", "chat_id": "6e4094947af4902cd252421aba9a077e8e4402dd", "models": "model-large___model-medium", "chat_item_ids": "0455df65c2d6bb821a9dc9108ac1d79964a0f571___4c28c86c4b22b3397691ce5cf27197fcf7e8fb2d___7326ff17cbfe7e3cb91b008cf0c496fcd17a1074___375b55d44af2c8c801992089c797df8e12605dfb___9231d8819ac96cc8a6c4b7780c301b796c3f8bf2", "text": "Chat_id: 6e4094947af4902cd252421aba9a077e8e4402dd\\n\\nChat_item_ids: 0455df65c2d6bb821a9dc9108ac1d79964a0f571___4c28c86c4b22b3397691ce5cf27197fcf7e8fb2d___7326ff17cbfe7e3cb91b008cf0c496fcd17a1074___375b55d44af2c8c801992089c797df8e12605dfb___9231d8819ac96cc8a6c4b7780c301b796c3f8bf2"}'
             ],
         )
         self.kili.graphql_client.execute.assert_called_with(*expected_parameters)
@@ -255,8 +255,8 @@ def test_process_json(self, *_):
                 },
                 {
                     "chat_id": "6e4094947af4902cd252421aba9a077e8e4402dd",
-                    "models": "model-large_model-medium",
-                    "chat_item_ids": "0455df65c2d6bb821a9dc9108ac1d79964a0f571_4c28c86c4b22b3397691ce5cf27197fcf7e8fb2d_7326ff17cbfe7e3cb91b008cf0c496fcd17a1074_375b55d44af2c8c801992089c797df8e12605dfb_9231d8819ac96cc8a6c4b7780c301b796c3f8bf2",
-                    "text": "Chat_id: 6e4094947af4902cd252421aba9a077e8e4402dd\n\nChat_item_ids: 0455df65c2d6bb821a9dc9108ac1d79964a0f571_4c28c86c4b22b3397691ce5cf27197fcf7e8fb2d_7326ff17cbfe7e3cb91b008cf0c496fcd17a1074_375b55d44af2c8c801992089c797df8e12605dfb_9231d8819ac96cc8a6c4b7780c301b796c3f8bf2",
+                    "models": "model-large___model-medium",
+                    "chat_item_ids": "0455df65c2d6bb821a9dc9108ac1d79964a0f571___4c28c86c4b22b3397691ce5cf27197fcf7e8fb2d___7326ff17cbfe7e3cb91b008cf0c496fcd17a1074___375b55d44af2c8c801992089c797df8e12605dfb___9231d8819ac96cc8a6c4b7780c301b796c3f8bf2",
+                    "text": "Chat_id: 6e4094947af4902cd252421aba9a077e8e4402dd\n\nChat_item_ids: 0455df65c2d6bb821a9dc9108ac1d79964a0f571___4c28c86c4b22b3397691ce5cf27197fcf7e8fb2d___7326ff17cbfe7e3cb91b008cf0c496fcd17a1074___375b55d44af2c8c801992089c797df8e12605dfb___9231d8819ac96cc8a6c4b7780c301b796c3f8bf2",
                 },
             )