arize-phoenix 4.7.2__py3-none-any.whl → 4.8.1__py3-none-any.whl

{arize_phoenix-4.7.2.dist-info → arize_phoenix-4.8.1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: arize-phoenix
-Version: 4.7.2
+Version: 4.8.1
 Summary: AI Observability and Evaluation
 Project-URL: Documentation, https://docs.arize.com/phoenix/
 Project-URL: Issues, https://github.com/Arize-ai/phoenix/issues

{arize_phoenix-4.7.2.dist-info → arize_phoenix-4.8.1.dist-info}/RECORD

@@ -5,7 +5,7 @@ phoenix/exceptions.py,sha256=n2L2KKuecrdflB9MsCdAYCiSEvGJptIsfRkXMoJle7A,169
 phoenix/py.typed,sha256=AbpHGcgLb-kRsJGnwFEktk7uzpZOCcBY74-YBdrKVGs,1
 phoenix/services.py,sha256=aTxhcOA1pZHB6U-B3TEcp6fqDF5oT0xCUvEUNMZVTUQ,5175
 phoenix/settings.py,sha256=cO-qgis_S27nHirTobYI9hHPfZH18R--WMmxNdsVUwc,273
-phoenix/version.py,sha256=cmBVf5yWtMWpLVrALEkcO90h-q8n53EpZSufTL15jyo,22
+phoenix/version.py,sha256=5LMFXHH6L0xgmxA0DlLrNZtG0vHgJ8uHcls9atNfDqI,22
 phoenix/core/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 phoenix/core/embedding_dimension.py,sha256=zKGbcvwOXgLf-yrJBpQyKtd-LEOPRKHnUToyAU8Owis,87
 phoenix/core/model.py,sha256=km_a--PBHOuA337ClRw9xqhOHhrUT6Rl9pz_zV0JYkQ,4843
@@ -16,7 +16,7 @@ phoenix/db/__init__.py,sha256=pDjEFXukHmJBM-1D8RjmXkvLsz85YWNxMQczt81ec3A,118
 phoenix/db/alembic.ini,sha256=p8DjVqGUs_tTx8oU56JP7qj-rMUebNFizItUSv_hPhs,3763
 phoenix/db/bulk_inserter.py,sha256=zbZGWZFDybKaGLGzpxgLwxAS5sC0_wXcvM0be4kUhh8,11286
 phoenix/db/engines.py,sha256=vLWaZlToMtDI7rJDxSidYkfOoojamxaZxaz8ND3zTus,4770
-phoenix/db/helpers.py,sha256=L2_jP1iIWpUREhKLYYb4_vf_6v_BiU1E73Z2PczGm6s,1589
+phoenix/db/helpers.py,sha256=mTBhPzdy_aU9gD7hNzUZJkAnV77ko5CdaXyoWH3snPA,2982
 phoenix/db/migrate.py,sha256=MuhtNWnR24riROvarvKfbRb4_D5xuQi6P760vBUKl1E,2270
 phoenix/db/models.py,sha256=7DBWbxY3cx3ve2P1I0kkDKXzlt04zEFJuRPJWsVpH-I,20422
 phoenix/db/insertion/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -71,6 +71,7 @@ phoenix/server/api/context.py,sha256=4jcy203Gtx38399FP21iU3HmFsq-50EKFJlX4IW2Los
 phoenix/server/api/interceptor.py,sha256=ykDnoC_apUd-llVli3m1CW18kNSIgjz2qZ6m5JmPDu8,1294
 phoenix/server/api/queries.py,sha256=wp5BlapuxDIoaQJm7mzG0dURfVxR32vXSJVC0JqG4_Y,19845
 phoenix/server/api/schema.py,sha256=BcxdqO5CSGqpKd-AAJHMjFlzaK9oJA8GJuxmMfcdjn4,434
+phoenix/server/api/utils.py,sha256=Y1lGu8J8r8BSBX9OzffgewI8QMziovbG-ePDvZrrwGI,949
 phoenix/server/api/dataloaders/__init__.py,sha256=qehXL37vGdw7v5PFs3kbZVIuhuzrVNVeZACDQjYpwyo,4847
 phoenix/server/api/dataloaders/average_experiment_run_latency.py,sha256=RiO0AKC6Y5byafsV0zTJEIOt8Nudjte73f1T78cBe1k,1817
 phoenix/server/api/dataloaders/dataset_example_revisions.py,sha256=Vpr5IEKSR4QnAVxE5NM7u92fPNgeHQV2ieYc6JakCj0,3788
@@ -85,7 +86,7 @@ phoenix/server/api/dataloaders/experiment_run_counts.py,sha256=wxHv08aZELJ91KTjH
 phoenix/server/api/dataloaders/experiment_sequence_number.py,sha256=Va1KuoHOd-wzvrlKykoV4kLRFW4JsJvGp_DUI4HYZX4,1631
 phoenix/server/api/dataloaders/latency_ms_quantile.py,sha256=pEc7QjB2iiNOQm_Fmo99F5O_DKOJWgGmcnT0OADJzYE,7423
 phoenix/server/api/dataloaders/min_start_or_max_end_times.py,sha256=IoFX5PtSpvQdMk_7-oB8TpIse3Q4PMxep4qKggkHpzo,2902
-phoenix/server/api/dataloaders/project_by_name.py,sha256=ziYp8fZGy2duD-f_oTqNLwbrGmBQFesqlTEHBDMtZlw,1170
+phoenix/server/api/dataloaders/project_by_name.py,sha256=W4q-ddeVnja2DNwgg5l8mV2GNADNJf1CTXNcJaBWdfE,1165
 phoenix/server/api/dataloaders/record_counts.py,sha256=mp3KlhwFw-Iy7o6bFxtJKC6B5kGinPIh5PxxNkrxf8o,4283
 phoenix/server/api/dataloaders/span_descendants.py,sha256=djqXXwupWaXkFdscoy-iILYijuxlqr8hJcv6JawsV6s,2162
 phoenix/server/api/dataloaders/span_evaluations.py,sha256=quXGyj_OYvHrcWfzXlYIv7C1OCZiL1K7cWU-YEAlsNA,1316
@@ -121,8 +122,8 @@ phoenix/server/api/input_types/TimeRange.py,sha256=yzx-gxj8mDeGLft1FzU_x1MVEgIG5
 phoenix/server/api/input_types/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 phoenix/server/api/mutations/__init__.py,sha256=Jcz-pM6SklhEGKcjA7AIB2WJXGhnxGsghrIq131kyDo,502
 phoenix/server/api/mutations/auth.py,sha256=vPRFoj7J6PV6QeODewG4K0PhoOebS5AfMRpbi_wuhyQ,311
-phoenix/server/api/mutations/dataset_mutations.py,sha256=Zp2sFWyGyubILUQboR6bafRWafsfeRO2ffUWnkLlfgI,22532
-phoenix/server/api/mutations/experiment_mutations.py,sha256=Fw_yEdITGJ6A33M5JZ-2YnBTDoBqZUUFON6vy8JoVjE,2569
+phoenix/server/api/mutations/dataset_mutations.py,sha256=CuKhxsYfvwVcdN_9EXhKxB6444BQfObzKzzyfAeg-n8,23199
+phoenix/server/api/mutations/experiment_mutations.py,sha256=vV2lbJ7ccXZqe-LY7nXx6QxWqhKQE4UNZAFcML-KQ8I,3011
 phoenix/server/api/mutations/export_events_mutations.py,sha256=t_wYBxaqvBJYRoHslh3Bmoxmwlzoy0u8SsBKWIKN5hE,4028
 phoenix/server/api/mutations/project_mutations.py,sha256=3SVDCZqxB0Iv60cOwBL8c-rY3QUUPs8PXbp-C_K1mWY,2267
 phoenix/server/api/openapi/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -130,9 +131,9 @@ phoenix/server/api/openapi/main.py,sha256=WY0pj3B7siQyyYqKyhqnzWC7P8MtEtiukOBUjG
 phoenix/server/api/openapi/schema.py,sha256=uuSYe1Ecu72aXRgTNjyMu-9ZPE13DAHJPKtedS-MsSs,451
 phoenix/server/api/routers/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 phoenix/server/api/routers/utils.py,sha256=M41BoH-fl37izhRuN2aX7lWm7jOC20A_3uClv9TVUUY,583
-phoenix/server/api/routers/v1/__init__.py,sha256=vvdpUa2LJPWEg8HbvDm_ANkBAwubPIFPbbHi7elOUws,2808
+phoenix/server/api/routers/v1/__init__.py,sha256=Ir5fsO6gQXW58HGm7s2sMUq0vya7mfcWneLcLJy6_q8,2895
 phoenix/server/api/routers/v1/dataset_examples.py,sha256=XfqOvDKF1oxb0pkeYfBycwwGt3LnSyyGdMLKC5VKoGQ,6690
-phoenix/server/api/routers/v1/datasets.py,sha256=f2gLG-geu-_wtEw4mKSzNWK2cFb5TYOyRL3tQ7Fl7Es,31544
+phoenix/server/api/routers/v1/datasets.py,sha256=r0WcNxF8SKVa3-4rrTIg4Andwr4NmRmW1ybpKuxR9qw,33639
 phoenix/server/api/routers/v1/evaluations.py,sha256=8g6P_e2BweV3RDU0esFmpkb0L5fCwonQPXiJ0y6HLwg,9126
 phoenix/server/api/routers/v1/experiment_evaluations.py,sha256=H_psVyuGUQImo0oxdEAKAMQ-oyVwkVIq5yaMHzHIiPc,5455
 phoenix/server/api/routers/v1/experiment_runs.py,sha256=jy4SynmzdtQMoUzlowmG6wsVU14SsLAzfcW4JOhXjeQ,8154
@@ -202,7 +203,7 @@ phoenix/server/static/apple-touch-icon-76x76.png,sha256=CT_xT12I0u2i0WU8JzBZBuOQ
 phoenix/server/static/apple-touch-icon.png,sha256=fOfpjqGpWYbJ0eAurKsyoZP1EAs6ZVooBJ_SGk2ZkDs,3801
 phoenix/server/static/favicon.ico,sha256=bY0vvCKRftemZfPShwZtE93DiiQdaYaozkPGwNFr6H8,34494
 phoenix/server/static/index.css,sha256=KKGpx4iwF91VGRm0YN-4cn8oC-oIqC6HecoPf0x3ZM8,1885
-phoenix/server/static/index.js,sha256=LmaCH-kjVuEBKvMhK3pOUxK12Wv26htH5juAzju9VWE,3524823
+phoenix/server/static/index.js,sha256=sUkm_imwx5uXpCQxQEPmX-axDr094HE1wGdBznF8J_E,3524949
 phoenix/server/static/modernizr.js,sha256=mvK-XtkNqjOral-QvzoqsyOMECXIMu5BQwSVN_wcU9c,2564
 phoenix/server/templates/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 phoenix/server/templates/index.html,sha256=S4z7qSoNSwnKFAH9r96AR-YJEyoKMd-VMWVlJ_IdzME,2039
@@ -221,9 +222,9 @@ phoenix/trace/otel.py,sha256=WA720jvRadiZBAKjsYoPyXzypHwbyEK2OZRVUwtbjB8,9976
 phoenix/trace/projects.py,sha256=2BwlNjFE-uwpqYtCu5YyBiYZk9wRPpM13vh3-Cv7GkA,2157
 phoenix/trace/schemas.py,sha256=Mjc6fD9OyeMnEk5wPPSbveqnNUYWK3p3BxpOvSGanHU,5950
 phoenix/trace/span_evaluations.py,sha256=GaADtJLi2njra4aYaie0BIwkSgdxPB_SNseglI4ykZA,13104
-phoenix/trace/span_json_decoder.py,sha256=IAFakPRqSMYxTPKYFMiXYxm7U-FipdN8_xbvapDS0Qc,3131
+phoenix/trace/span_json_decoder.py,sha256=W6T1U0xndE2a-85O1QalJmStJK1zF1IoSU97h1RoWeY,3231
 phoenix/trace/span_json_encoder.py,sha256=tzSCIQJbeFBm33K68G8A5M12n_86tCDyuU0WAobxEz4,2010
-phoenix/trace/trace_dataset.py,sha256=CI9oma7sJTggUxvNHQR-uGIqrAvcSvkcDfyUBOeWjxo,13848
+phoenix/trace/trace_dataset.py,sha256=9qDs8T3QQOTN9fHZzyeJlAvR-yW7tkq1OvV0xWILgzA,13917
 phoenix/trace/utils.py,sha256=7LurVGXn245cjj4MJsc7v6jq4DSJkpK6YGBfIaSywuw,1307
 phoenix/trace/dsl/README.md,sha256=ihmP9zGUC5V-TDbzKla76LuyDqPDQIBUH2BORwxNI68,2902
 phoenix/trace/dsl/__init__.py,sha256=WIQIjJg362XD3s50OsPJJ0xbDsGp41bSv7vDllLrPuA,144
@@ -247,8 +248,8 @@ phoenix/utilities/logging.py,sha256=lDXd6EGaamBNcQxL4vP1au9-i_SXe0OraUDiJOcszSw,
 phoenix/utilities/project.py,sha256=qWsvKnG1oKhOFUowXf9qiOL2ia7jaFe_ijFFHEt8GJo,431
 phoenix/utilities/re.py,sha256=PDve_OLjRTM8yQQJHC8-n3HdIONi7aNils3ZKRZ5uBM,2045
 phoenix/utilities/span_store.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-arize_phoenix-4.7.2.dist-info/METADATA,sha256=m2l2uSMpqkXD-NpYdOBqsKlNF5jMUoLoSjnxo9is9h0,11451
-arize_phoenix-4.7.2.dist-info/WHEEL,sha256=1yFddiXMmvYK7QYTqtRNtX66WJ0Mz8PYEiEUoOUUxRY,87
-arize_phoenix-4.7.2.dist-info/licenses/IP_NOTICE,sha256=JBqyyCYYxGDfzQ0TtsQgjts41IJoa-hiwDrBjCb9gHM,469
-arize_phoenix-4.7.2.dist-info/licenses/LICENSE,sha256=HFkW9REuMOkvKRACuwLPT0hRydHb3zNg-fdFt94td18,3794
-arize_phoenix-4.7.2.dist-info/RECORD,,
+arize_phoenix-4.8.1.dist-info/METADATA,sha256=cr_RMue_eqzBxlqGELzqha7zvv0EF_J9qfc7grNpekg,11451
+arize_phoenix-4.8.1.dist-info/WHEEL,sha256=1yFddiXMmvYK7QYTqtRNtX66WJ0Mz8PYEiEUoOUUxRY,87
+arize_phoenix-4.8.1.dist-info/licenses/IP_NOTICE,sha256=JBqyyCYYxGDfzQ0TtsQgjts41IJoa-hiwDrBjCb9gHM,469
+arize_phoenix-4.8.1.dist-info/licenses/LICENSE,sha256=HFkW9REuMOkvKRACuwLPT0hRydHb3zNg-fdFt94td18,3794
+arize_phoenix-4.8.1.dist-info/RECORD,,

phoenix/db/helpers.py

@@ -1,12 +1,12 @@
 from enum import Enum
-from typing import Any
+from typing import Any, Optional, Tuple
 
 from openinference.semconv.trace import (
     OpenInferenceSpanKindValues,
     RerankerAttributes,
     SpanAttributes,
 )
-from sqlalchemy import Integer, SQLColumnExpression, case, func
+from sqlalchemy import Integer, Select, SQLColumnExpression, case, distinct, func, select
 from typing_extensions import assert_never
 
 from phoenix.db import models
@@ -45,3 +45,38 @@ def num_docs_col(dialect: SupportedSQLDialect) -> SQLColumnExpression[Integer]:
 
 _RETRIEVAL_DOCUMENTS = SpanAttributes.RETRIEVAL_DOCUMENTS.split(".")
 _RERANKER_OUTPUT_DOCUMENTS = RerankerAttributes.RERANKER_OUTPUT_DOCUMENTS.split(".")
+
+
+def get_eval_trace_ids_for_datasets(*dataset_ids: int) -> Select[Tuple[Optional[str]]]:
+    return (
+        select(distinct(models.ExperimentRunAnnotation.trace_id))
+        .join(models.ExperimentRun)
+        .join_from(models.ExperimentRun, models.Experiment)
+        .where(models.Experiment.dataset_id.in_(set(dataset_ids)))
+        .where(models.ExperimentRunAnnotation.trace_id.isnot(None))
+    )
+
+
+def get_project_names_for_datasets(*dataset_ids: int) -> Select[Tuple[Optional[str]]]:
+    return (
+        select(distinct(models.Experiment.project_name))
+        .where(models.Experiment.dataset_id.in_(set(dataset_ids)))
+        .where(models.Experiment.project_name.isnot(None))
+    )
+
+
+def get_eval_trace_ids_for_experiments(*experiment_ids: int) -> Select[Tuple[Optional[str]]]:
+    return (
+        select(distinct(models.ExperimentRunAnnotation.trace_id))
+        .join(models.ExperimentRun)
+        .where(models.ExperimentRun.experiment_id.in_(set(experiment_ids)))
+        .where(models.ExperimentRunAnnotation.trace_id.isnot(None))
+    )
+
+
+def get_project_names_for_experiments(*experiment_ids: int) -> Select[Tuple[Optional[str]]]:
+    return (
+        select(distinct(models.Experiment.project_name))
+        .where(models.Experiment.id.in_(set(experiment_ids)))
+        .where(models.Experiment.project_name.isnot(None))
+    )

phoenix/server/api/dataloaders/project_by_name.py

@@ -28,4 +28,4 @@ class ProjectByNameDataLoader(DataLoader[Key, Result]):
             async for project in data:
                 projects_by_name[project.name] = project
 
-        return [projects_by_name[project_name] for project_name in project_names]
+        return [projects_by_name.get(project_name) for project_name in keys]

phoenix/server/api/mutations/dataset_mutations.py

@@ -1,3 +1,4 @@
+import asyncio
 from datetime import datetime
 from typing import Any, Dict
 
@@ -10,6 +11,7 @@ from strawberry import UNSET
 from strawberry.types import Info
 
 from phoenix.db import models
+from phoenix.db.helpers import get_eval_trace_ids_for_datasets, get_project_names_for_datasets
 from phoenix.server.api.context import Context
 from phoenix.server.api.helpers.dataset_helpers import (
     get_dataset_example_input,
@@ -30,6 +32,7 @@ from phoenix.server.api.types.Dataset import Dataset, to_gql_dataset
 from phoenix.server.api.types.DatasetExample import DatasetExample
 from phoenix.server.api.types.node import from_global_id_with_expected_type
 from phoenix.server.api.types.Span import Span
+from phoenix.server.api.utils import delete_projects, delete_traces
 
 
 @strawberry.type
@@ -274,21 +277,28 @@ class DatasetMutationMixin:
         info: Info[Context, None],
         input: DeleteDatasetInput,
     ) -> DatasetMutationPayload:
-        dataset_id = input.dataset_id
-        dataset_rowid = from_global_id_with_expected_type(
-            global_id=dataset_id, expected_type_name=Dataset.__name__
+        try:
+            dataset_id = from_global_id_with_expected_type(
+                global_id=input.dataset_id,
+                expected_type_name=Dataset.__name__,
+            )
+        except ValueError:
+            raise ValueError(f"Unknown dataset: {input.dataset_id}")
+        project_names_stmt = get_project_names_for_datasets(dataset_id)
+        eval_trace_ids_stmt = get_eval_trace_ids_for_datasets(dataset_id)
+        stmt = (
+            delete(models.Dataset).where(models.Dataset.id == dataset_id).returning(models.Dataset)
         )
-
         async with info.context.db() as session:
-            delete_result = await session.execute(
-                delete(models.Dataset)
-                .where(models.Dataset.id == dataset_rowid)
-                .returning(models.Dataset)
-            )
-            if not (datasets := delete_result.first()):
-                raise ValueError(f"Unknown dataset: {dataset_id}")
-
-        dataset = datasets[0]
+            project_names = await session.scalars(project_names_stmt)
+            eval_trace_ids = await session.scalars(eval_trace_ids_stmt)
+            if not (dataset := await session.scalar(stmt)):
+                raise ValueError(f"Unknown dataset: {input.dataset_id}")
+        await asyncio.gather(
+            delete_projects(info.context.db, *project_names),
+            delete_traces(info.context.db, *eval_trace_ids),
+            return_exceptions=True,
+        )
         return DatasetMutationPayload(dataset=to_gql_dataset(dataset))
 
     @strawberry.mutation(permission_classes=[IsAuthenticated])  # type: ignore

phoenix/server/api/mutations/experiment_mutations.py

@@ -1,3 +1,4 @@
+import asyncio
 from typing import List
 
 import strawberry
@@ -6,11 +7,13 @@ from strawberry.relay import GlobalID
 from strawberry.types import Info
 
 from phoenix.db import models
+from phoenix.db.helpers import get_eval_trace_ids_for_experiments, get_project_names_for_experiments
 from phoenix.server.api.context import Context
 from phoenix.server.api.input_types.DeleteExperimentsInput import DeleteExperimentsInput
 from phoenix.server.api.mutations.auth import IsAuthenticated
 from phoenix.server.api.types.Experiment import Experiment, to_gql_experiment
 from phoenix.server.api.types.node import from_global_id_with_expected_type
+from phoenix.server.api.utils import delete_projects, delete_traces
 
 
 @strawberry.type
@@ -30,7 +33,11 @@ class ExperimentMutationMixin:
             from_global_id_with_expected_type(experiment_id, Experiment.__name__)
             for experiment_id in input.experiment_ids
         ]
+        project_names_stmt = get_project_names_for_experiments(*experiment_ids)
+        eval_trace_ids_stmt = get_eval_trace_ids_for_experiments(*experiment_ids)
         async with info.context.db() as session:
+            project_names = await session.scalars(project_names_stmt)
+            eval_trace_ids = await session.scalars(eval_trace_ids_stmt)
             savepoint = await session.begin_nested()
             experiments = {
                 experiment.id: experiment
@@ -54,10 +61,11 @@ class ExperimentMutationMixin:
                         ]
                     )
                 )
-            if project_names := set(filter(bool, (e.project_name for e in experiments.values()))):
-                await session.execute(
-                    delete(models.Project).where(models.Project.name.in_(project_names))
-                )
+        await asyncio.gather(
+            delete_projects(info.context.db, *project_names),
+            delete_traces(info.context.db, *eval_trace_ids),
+            return_exceptions=True,
+        )
         return ExperimentMutationPayload(
             experiments=[
                 to_gql_experiment(experiments[experiment_id]) for experiment_id in experiment_ids

phoenix/server/api/routers/v1/__init__.py

@@ -44,6 +44,7 @@ V1_ROUTES = [
     Route("/v1/spans", spans.get_spans_handler, methods=["GET"]),
     Route("/v1/datasets/upload", datasets.post_datasets_upload, methods=["POST"]),
     Route("/v1/datasets", datasets.list_datasets, methods=["GET"]),
+    Route("/v1/datasets/{id:str}", datasets.delete_dataset_by_id, methods=["DELETE"]),
     Route("/v1/datasets/{id:str}", datasets.get_dataset_by_id, methods=["GET"]),
     Route("/v1/datasets/{id:str}/csv", datasets.get_dataset_csv, methods=["GET"]),
     Route(

phoenix/server/api/routers/v1/datasets.py

@@ -26,13 +26,15 @@ from typing import (
 
 import pandas as pd
 import pyarrow as pa
-from sqlalchemy import and_, func, select
+from sqlalchemy import and_, delete, func, select
 from sqlalchemy.ext.asyncio import AsyncSession
+from starlette.background import BackgroundTasks
 from starlette.concurrency import run_in_threadpool
 from starlette.datastructures import FormData, UploadFile
 from starlette.requests import Request
 from starlette.responses import JSONResponse, Response
 from starlette.status import (
+    HTTP_204_NO_CONTENT,
     HTTP_404_NOT_FOUND,
     HTTP_409_CONFLICT,
     HTTP_422_UNPROCESSABLE_ENTITY,
@@ -42,6 +44,7 @@ from strawberry.relay import GlobalID
 from typing_extensions import TypeAlias, assert_never
 
 from phoenix.db import models
+from phoenix.db.helpers import get_eval_trace_ids_for_datasets, get_project_names_for_datasets
 from phoenix.db.insertion.dataset import (
     DatasetAction,
     DatasetExampleAdditionEvent,
@@ -52,6 +55,7 @@ from phoenix.server.api.types.Dataset import Dataset
 from phoenix.server.api.types.DatasetExample import DatasetExample
 from phoenix.server.api.types.DatasetVersion import DatasetVersion
 from phoenix.server.api.types.node import from_global_id_with_expected_type
+from phoenix.server.api.utils import delete_projects, delete_traces
 
 logger = logging.getLogger(__name__)
 
@@ -163,6 +167,60 @@ async def list_datasets(request: Request) -> Response:
         return JSONResponse(content={"next_cursor": next_cursor, "data": data})
 
 
+async def delete_dataset_by_id(request: Request) -> Response:
+    """
+    summary: Delete dataset by ID
+    operationId: deleteDatasetById
+    tags:
+      - datasets
+    parameters:
+      - in: path
+        name: id
+        required: true
+        schema:
+          type: string
+    responses:
+      204:
+        description: Success
+      403:
+        description: Forbidden
+      404:
+        description: Dataset not found
+      422:
+        description: Dataset ID is invalid
+    """
+    if id_ := request.path_params.get("id"):
+        try:
+            dataset_id = from_global_id_with_expected_type(
+                GlobalID.from_id(id_),
+                Dataset.__name__,
+            )
+        except ValueError:
+            return Response(
+                content=f"Invalid Dataset ID: {id_}",
+                status_code=HTTP_422_UNPROCESSABLE_ENTITY,
+            )
+    else:
+        return Response(
+            content="Missing Dataset ID",
+            status_code=HTTP_422_UNPROCESSABLE_ENTITY,
+        )
+    project_names_stmt = get_project_names_for_datasets(dataset_id)
+    eval_trace_ids_stmt = get_eval_trace_ids_for_datasets(dataset_id)
+    stmt = (
+        delete(models.Dataset).where(models.Dataset.id == dataset_id).returning(models.Dataset.id)
+    )
+    async with request.app.state.db() as session:
+        project_names = await session.scalars(project_names_stmt)
+        eval_trace_ids = await session.scalars(eval_trace_ids_stmt)
+        if (await session.scalar(stmt)) is None:
+            return Response(content="Dataset does not exist", status_code=HTTP_404_NOT_FOUND)
+    tasks = BackgroundTasks()
+    tasks.add_task(delete_projects, request.app.state.db, *project_names)
+    tasks.add_task(delete_traces, request.app.state.db, *eval_trace_ids)
+    return Response(status_code=HTTP_204_NO_CONTENT, background=tasks)
+
+
 async def get_dataset_by_id(request: Request) -> Response:
     """
     summary: Get dataset by ID

phoenix/server/api/utils.py

@@ -0,0 +1,36 @@
+from typing import AsyncContextManager, Callable, List
+
+from sqlalchemy import delete
+from sqlalchemy.ext.asyncio import AsyncSession
+
+from phoenix.db import models
+
+
+async def delete_projects(
+    db: Callable[[], AsyncContextManager[AsyncSession]],
+    *project_names: str,
+) -> List[int]:
+    if not project_names:
+        return []
+    stmt = (
+        delete(models.Project)
+        .where(models.Project.name.in_(set(project_names)))
+        .returning(models.Project.id)
+    )
+    async with db() as session:
+        return list(await session.scalars(stmt))
+
+
+async def delete_traces(
+    db: Callable[[], AsyncContextManager[AsyncSession]],
+    *trace_ids: str,
+) -> List[int]:
+    if not trace_ids:
+        return []
+    stmt = (
+        delete(models.Trace)
+        .where(models.Trace.trace_id.in_(set(trace_ids)))
+        .returning(models.Trace.id)
+    )
+    async with db() as session:
+        return list(await session.scalars(stmt))

phoenix/server/static/index.js

@@ -7750,7 +7750,7 @@ fragment DatasetsTable_datasets on Query {
     __typename
   }
 }
-`}}}();bat.hash="d98c682a2c908030c2f772209a4a25b8";var mat=bat;var Q0=P(ie()),Iat=P(wt());function pat(t){let{datasetId:e,datasetName:n,onDatasetDelete:i,onDatasetDeleteError:r}=t,[a,o]=(0,Q0.useState)(null),[l,s]=(0,Iat.useMutation)(mat),c=(0,Q0.useCallback)(()=>{(0,Q0.startTransition)(()=>{l({variables:{datasetId:e},onCompleted:()=>{i()},onError:g=>{r(g)}})})},[l,e,i,r]),u=(0,Q0.useCallback)(()=>{o(B(Mn,{size:"S",title:"Delete Dataset",children:[h(De,{padding:"size-200",children:h(me,{color:"danger",children:`Are you sure you want to delete dataset ${n}? This cannot be undone.`})}),h(De,{paddingEnd:"size-200",paddingTop:"size-100",paddingBottom:"size-100",borderTopColor:"light",borderTopWidth:"thin",children:h(ge,{direction:"row",justifyContent:"end",children:h(At,{variant:"danger",onClick:()=>{c(),o(null)},children:"Delete Dataset"})})})]}))},[c,n]);return B("div",{onClick:g=>{g.preventDefault(),g.stopPropagation()},children:[h(Gc,{align:"end",buttonSize:"compact",isDisabled:s,onAction:g=>{switch(g){case"deleteDataset":u();break}},children:h(Ht,{children:B(ge,{direction:"row",gap:"size-75",justifyContent:"start",alignItems:"center",children:[h(ke,{svg:h(Pe.TrashOutline,{})}),h(me,{children:"Delete"})]})},"deleteDataset")}),h(Fn,{type:"modal",isDismissable:!0,onDismiss:()=>o(null),children:a})]})}var hat=100;function cGn(t){let e=t.id;if(e!=="createdAt"&&e!=="name")throw new Error("Invalid sort column");return{col:e,dir:t.desc?"desc":"asc"}}function Cat(t){let[e,n]=(0,cd.useState)([]),i=(0,cd.useRef)(null),r=Zn(),a=Zo(),o=dm(),{data:l,loadNext:s,hasNext:c,isLoadingNext:u,refetch:g}=(0,fat.usePaginationFragment)(dat,t.query),d=(0,cd.useMemo)(()=>l.datasets.edges.map(f=>f.node),[l]),b=cd.default.useCallback(f=>{if(f){let{scrollHeight:C,scrollTop:v,clientHeight:x}=f;C-v-x<300&&!u&&c&&s(hat)}},[c,u,s]),m=Gr({columns:[{header:"name",accessorKey:"name",cell:({row:f})=>{let v=f.original.experimentCount>0?`${f.original.id}/experiments`:`${f.original.id}/examples`;return h(_l,{to:v,children:f.original.name})}},{header:"description",accessorKey:"description",enableSorting:!1},{header:"created at",accessorKey:"createdAt",cell:sd},{header:"example count",accessorKey:"exampleCount",enableSorting:!1,meta:{textAlign:"right"}},{header:"experiment count",accessorKey:"experimentCount",enableSorting:!1,meta:{textAlign:"right"}},{header:"",id:"actions",enableSorting:!1,size:10,cell:({row:f})=>h(pat,{datasetId:f.original.id,datasetName:f.original.name,onDatasetDelete:()=>{a({title:"Dataset deleted",message:`${f.original.name} has been successfully deleted.`}),g({},{fetchPolicy:"store-and-network"})},onDatasetDeleteError:C=>{o({title:"Dataset deletion failed",message:C.message})}})}],data:d,state:{sorting:e},getCoreRowModel:Wr(),getSortedRowModel:Zg(),onSortingChange:n,manualSorting:!0});(0,cd.useEffect)(()=>{let f=e[0];(0,cd.startTransition)(()=>{g({sort:f?cGn(f):{col:"createdAt",dir:"desc"},after:null,first:hat},{fetchPolicy:"store-and-network"})})},[e,g]);let I=m.getRowModel().rows,p=I.length===0;return h("div",{css:$`
+`}}}();bat.hash="d98c682a2c908030c2f772209a4a25b8";var mat=bat;var Q0=P(ie()),Iat=P(wt());function pat(t){let{datasetId:e,datasetName:n,onDatasetDelete:i,onDatasetDeleteError:r}=t,[a,o]=(0,Q0.useState)(null),[l,s]=(0,Iat.useMutation)(mat),c=(0,Q0.useCallback)(()=>{(0,Q0.startTransition)(()=>{l({variables:{datasetId:e},onCompleted:()=>{i()},onError:g=>{r(g)}})})},[l,e,i,r]),u=(0,Q0.useCallback)(()=>{o(B(Mn,{size:"S",title:"Delete Dataset",children:[h(De,{padding:"size-200",children:h(me,{color:"danger",children:`Are you sure you want to delete dataset ${n}? This will also delete all associated experiments and traces, and it cannot be undone.`})}),h(De,{paddingEnd:"size-200",paddingTop:"size-100",paddingBottom:"size-100",borderTopColor:"light",borderTopWidth:"thin",children:h(ge,{direction:"row",justifyContent:"end",children:h(At,{variant:"danger",onClick:()=>{c(),o(null)},children:"Delete Dataset"})})})]}))},[c,n]);return B("div",{onClick:g=>{g.preventDefault(),g.stopPropagation()},children:[h(Gc,{align:"end",buttonSize:"compact",isDisabled:s,onAction:g=>{switch(g){case"deleteDataset":u();break}},children:h(Ht,{children:B(ge,{direction:"row",gap:"size-75",justifyContent:"start",alignItems:"center",children:[h(ke,{svg:h(Pe.TrashOutline,{})}),h(me,{children:"Delete"})]})},"deleteDataset")}),h(Fn,{type:"modal",isDismissable:!0,onDismiss:()=>o(null),children:a})]})}var hat=100;function cGn(t){let e=t.id;if(e!=="createdAt"&&e!=="name")throw new Error("Invalid sort column");return{col:e,dir:t.desc?"desc":"asc"}}function Cat(t){let[e,n]=(0,cd.useState)([]),i=(0,cd.useRef)(null),r=Zn(),a=Zo(),o=dm(),{data:l,loadNext:s,hasNext:c,isLoadingNext:u,refetch:g}=(0,fat.usePaginationFragment)(dat,t.query),d=(0,cd.useMemo)(()=>l.datasets.edges.map(f=>f.node),[l]),b=cd.default.useCallback(f=>{if(f){let{scrollHeight:C,scrollTop:v,clientHeight:x}=f;C-v-x<300&&!u&&c&&s(hat)}},[c,u,s]),m=Gr({columns:[{header:"name",accessorKey:"name",cell:({row:f})=>{let v=f.original.experimentCount>0?`${f.original.id}/experiments`:`${f.original.id}/examples`;return h(_l,{to:v,children:f.original.name})}},{header:"description",accessorKey:"description",enableSorting:!1},{header:"created at",accessorKey:"createdAt",cell:sd},{header:"example count",accessorKey:"exampleCount",enableSorting:!1,meta:{textAlign:"right"}},{header:"experiment count",accessorKey:"experimentCount",enableSorting:!1,meta:{textAlign:"right"}},{header:"",id:"actions",enableSorting:!1,size:10,cell:({row:f})=>h(pat,{datasetId:f.original.id,datasetName:f.original.name,onDatasetDelete:()=>{a({title:"Dataset deleted",message:`${f.original.name} has been successfully deleted.`}),g({},{fetchPolicy:"store-and-network"})},onDatasetDeleteError:C=>{o({title:"Dataset deletion failed",message:C.message})}})}],data:d,state:{sorting:e},getCoreRowModel:Wr(),getSortedRowModel:Zg(),onSortingChange:n,manualSorting:!0});(0,cd.useEffect)(()=>{let f=e[0];(0,cd.startTransition)(()=>{g({sort:f?cGn(f):{col:"createdAt",dir:"desc"},after:null,first:hat},{fetchPolicy:"store-and-network"})})},[e,g]);let I=m.getRowModel().rows,p=I.length===0;return h("div",{css:$`
         flex: 1 1 auto;
         overflow: auto;
       `,onScroll:f=>b(f.target),ref:i,children:B("table",{css:wg,children:[h("thead",{children:m.getHeaderGroups().map(f=>h("tr",{children:f.headers.map(C=>h("th",{colSpan:C.colSpan,children:C.isPlaceholder?null:B("div",{className:C.column.getCanSort()?"cursor-pointer":"",onClick:C.column.getToggleSortingHandler(),style:{textAlign:C.column.columnDef.meta?.textAlign},children:[Hn(C.column.columnDef.header,C.getContext()),C.column.getIsSorted()?h(ke,{className:"sort-icon",svg:C.column.getIsSorted()==="asc"?h(Pe.ArrowUpFilled,{}):h(Pe.ArrowDownFilled,{})}):null]})},C.id))},f.id))}),p?h(Ul,{}):h("tbody",{children:I.map(f=>h("tr",{onClick:()=>{let v=f.original.experimentCount>0?`${f.original.id}/experiments`:`${f.original.id}/examples`;r(v)},children:f.getVisibleCells().map(C=>h("td",{align:C.column.columnDef.meta?.textAlign,children:Hn(C.column.columnDef.cell,C.getContext())},C.id))},f.id))})]})})}function yat(){return h(K0.Suspense,{fallback:h(Qr,{}),children:h(uGn,{})})}function uGn(){let[t,e]=(0,K0.useState)(0),n=(0,Aat.useLazyLoadQuery)(oat,{},{fetchKey:t}),i=(0,K0.useCallback)(()=>{e(r=>r+1)},[e]);return B(ge,{direction:"column",height:"100%",children:[h(De,{padding:"size-200",borderBottomWidth:"thin",borderBottomColor:"dark",flex:"none",children:B(ge,{direction:"row",justifyContent:"space-between",children:[h(dn,{level:1,children:"Datasets"}),h(gGn,{onDatasetCreated:i})]})}),h(Cat,{query:n})]})}function gGn({onDatasetCreated:t}){let e=Zn(),n=Zo(),i=dm(),[r,a]=(0,K0.useState)(null),o=()=>{a(h(Mn,{size:"S",title:"New Dataset",children:h(rY,{onDatasetCreated:s=>{n({title:"Dataset created",message:`${s.name} has been successfully created.`,action:{text:"Go to Dataset",onClick:()=>{e(`/datasets/${s.id}`)}}}),a(null),t()},onDatasetCreateError:s=>{i({title:"Dataset creation failed",message:s.message})}})}))},l=()=>{a(h(Mn,{size:"M",title:"New Dataset from CSV",children:h(lat,{onDatasetCreated:s=>{n({title:"Dataset created",message:`${s.name} has been successfully created.`,action:{text:"Go to Dataset",onClick:()=>{e(`/datasets/${s.id}`)}}}),a(null),t()},onDatasetCreateError:s=>{i({title:"Dataset creation failed",message:s.message})}})}))};return B($t,{children:[B(Gc,{buttonText:"Create Dataset",align:"end",icon:h(ke,{svg:h(Pe.DatabaseOutline,{})}),onAction:s=>{switch(s){case"newDataset":o();break;case"datasetFromCSV":l();break}},children:[h(Ht,{children:"New Dataset"},"newDataset"),h(Ht,{children:"Dataset from CSV"},"datasetFromCSV")]}),h(Fn,{type:"modal",isDismissable:!0,onDismiss:()=>a(null),children:r})]})}var tA=P(ie());var uY=P(ie());var vat=function(){var t=[{defaultValue:null,kind:"LocalArgument",name:"datasetId"}],e=[{kind:"Variable",name:"id",variableName:"datasetId"}],n={alias:null,args:null,kind:"ScalarField",name:"id",storageKey:null},i={kind:"InlineFragment",selections:[{alias:"latestVersions",args:[{kind:"Literal",name:"first",value:1},{kind:"Literal",name:"sort",value:{col:"createdAt",dir:"desc"}}],concreteType:"DatasetVersionConnection",kind:"LinkedField",name:"versions",plural:!1,selections:[{alias:null,args:null,concreteType:"DatasetVersionEdge",kind:"LinkedField",name:"edges",plural:!0,selections:[{alias:"version",args:null,concreteType:"DatasetVersion",kind:"LinkedField",name:"node",plural:!1,selections:[n,{alias:null,args:null,kind:"ScalarField",name:"description",storageKey:null},{alias:null,args:null,kind:"ScalarField",name:"createdAt",storageKey:null}],storageKey:null}],storageKey:null}],storageKey:'versions(first:1,sort:{"col":"createdAt","dir":"desc"})'}],type:"Dataset",abstractKey:null};return{fragment:{argumentDefinitions:t,kind:"Fragment",metadata:null,name:"datasetStore_latestVersionQuery",selections:[{alias:"dataset",args:e,concreteType:null,kind:"LinkedField",name:"node",plural:!1,selections:[n,i],storageKey:null}],type:"Query",abstractKey:null},kind:"Request",operation:{argumentDefinitions:t,kind:"Operation",name:"datasetStore_latestVersionQuery",selections:[{alias:"dataset",args:e,concreteType:null,kind:"LinkedField",name:"node",plural:!1,selections:[{alias:null,args:null,kind:"ScalarField",name:"__typename",storageKey:null},n,i],storageKey:null}]},params:{cacheID:"d7075240e4dae8871997e57616c05302",id:null,metadata:{},name:"datasetStore_latestVersionQuery",operationKind:"query",text:`query datasetStore_latestVersionQuery(
@@ -8152,7 +8152,7 @@ fragment ExperimentsTableFragment on Dataset {
     __typename
   }
 }
-`}}}();Vot.hash="4ab310746eb51c6a3fcb799d2c972762";var Eot=Vot;var a3=P(ie()),Jot=P(wt());function Hot(t){let e=Zn(),[n,i]=(0,a3.useState)(null),[r,a]=(0,Jot.useMutation)(Eot),{datasetId:o,selectedExperiments:l,onClearSelection:s,onExperimentsDeleted:c}=t,u=l.length!==1,g=Zo(),d=dm(),b=(0,a3.useCallback)(()=>{r({variables:{input:{experimentIds:l.map(I=>I.id)}},onCompleted:()=>{g({title:"Examples Deleted",message:`${l.length} experiment${u?"s":""} have been deleted.`}),c(),s()},onError:I=>{d({title:"An error occurred",message:`Failed to delete examples: ${I.message}`})}})},[r,u,d,g,s,c,l]),m=(0,a3.useCallback)(()=>{i(B(Mn,{size:"S",title:"Delete Experiments",children:[h(De,{padding:"size-200",children:h(me,{color:"danger",children:"Are you sure you want to delete these experiments? This cannot be undone."})}),h(De,{paddingEnd:"size-200",paddingTop:"size-100",paddingBottom:"size-100",borderTopColor:"light",borderTopWidth:"thin",children:h(ge,{direction:"row",justifyContent:"end",children:h(At,{variant:"danger",onClick:()=>{b(),i(null)},children:"Delete Experiments"})})})]}))},[b]);return B("div",{css:$`
+`}}}();Vot.hash="4ab310746eb51c6a3fcb799d2c972762";var Eot=Vot;var a3=P(ie()),Jot=P(wt());function Hot(t){let e=Zn(),[n,i]=(0,a3.useState)(null),[r,a]=(0,Jot.useMutation)(Eot),{datasetId:o,selectedExperiments:l,onClearSelection:s,onExperimentsDeleted:c}=t,u=l.length!==1,g=Zo(),d=dm(),b=(0,a3.useCallback)(()=>{r({variables:{input:{experimentIds:l.map(I=>I.id)}},onCompleted:()=>{g({title:"Examples Deleted",message:`${l.length} experiment${u?"s":""} have been deleted.`}),c(),s()},onError:I=>{d({title:"An error occurred",message:`Failed to delete examples: ${I.message}`})}})},[r,u,d,g,s,c,l]),m=(0,a3.useCallback)(()=>{i(B(Mn,{size:"S",title:"Delete Experiments",children:[h(De,{padding:"size-200",children:h(me,{color:"danger",children:"Are you sure you want to delete these experiments? This will also delete all associated annotations and traces, and it cannot be undone."})}),h(De,{paddingEnd:"size-200",paddingTop:"size-100",paddingBottom:"size-100",borderTopColor:"light",borderTopWidth:"thin",children:h(ge,{direction:"row",justifyContent:"end",children:h(At,{variant:"danger",onClick:()=>{b(),i(null)},children:"Delete Experiments"})})})]}))},[b]);return B("div",{css:$`
         position: absolute;
         bottom: var(--ac-global-dimension-size-400);
         left: 50%;

phoenix/trace/span_json_decoder.py

@@ -80,7 +80,9 @@ def json_to_span(data: Dict[str, Any]) -> Any:
                 attributes=event.get("attributes") or {},
                 timestamp=datetime.fromisoformat(event["timestamp"]),
             )
-            for event in data["events"]
+            for event in (
+                data["events"] if isinstance(data["events"], list) else json.loads(data["events"])
+            )
         ]
         data["conversation"] = (
             SpanConversationAttributes(**data["conversation"])

phoenix/trace/trace_dataset.py

@@ -5,6 +5,7 @@ from typing import Any, Iterable, Iterator, List, Optional, Tuple, Union, cast
 from uuid import UUID, uuid4
 from warnings import warn
 
+import numpy as np
 import pandas as pd
 from openinference.semconv.trace import (
     DocumentAttributes,
@@ -60,6 +61,7 @@ def normalize_dataframe(dataframe: DataFrame) -> "DataFrame":
     # Convert the start and end times to datetime
     dataframe["start_time"] = normalize_timestamps(dataframe["start_time"])
     dataframe["end_time"] = normalize_timestamps(dataframe["end_time"])
+    dataframe = dataframe.replace({np.nan: None})
     return dataframe
 
 

phoenix/version.py

@@ -1 +1 @@
-__version__ = "4.7.2"
+__version__ = "4.8.1"