arize-phoenix 9.3.0__tar.gz → 9.5.0__tar.gz

{arize_phoenix-9.3.0 → arize_phoenix-9.5.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: arize-phoenix
-Version: 9.3.0
+Version: 9.5.0
 Summary: AI Observability and Evaluation
 Project-URL: Documentation, https://docs.arize.com/phoenix/
 Project-URL: Issues, https://github.com/Arize-ai/phoenix/issues
@@ -49,7 +49,7 @@ Requires-Dist: scipy
 Requires-Dist: sqlalchemy[asyncio]<3,>=2.0.4
 Requires-Dist: sqlean-py>=3.45.1
 Requires-Dist: starlette
-Requires-Dist: strawberry-graphql==0.267.0
+Requires-Dist: strawberry-graphql==0.269.0
 Requires-Dist: tqdm
 Requires-Dist: typing-extensions>=4.6
 Requires-Dist: uvicorn
@@ -71,7 +71,7 @@ Requires-Dist: opentelemetry-sdk; extra == 'container'
 Requires-Dist: opentelemetry-semantic-conventions; extra == 'container'
 Requires-Dist: prometheus-client; extra == 'container'
 Requires-Dist: py-grpc-prometheus; extra == 'container'
-Requires-Dist: strawberry-graphql[opentelemetry]==0.267.0; extra == 'container'
+Requires-Dist: strawberry-graphql[opentelemetry]==0.269.0; extra == 'container'
 Requires-Dist: umap-learn; extra == 'container'
 Requires-Dist: uvloop; (platform_system != 'Windows') and extra == 'container'
 Provides-Extra: dev
@@ -102,7 +102,7 @@ Requires-Dist: pytest-postgresql; extra == 'dev'
 Requires-Dist: pytest-xdist; extra == 'dev'
 Requires-Dist: pytest==8.3.3; extra == 'dev'
 Requires-Dist: ruff==0.6.9; extra == 'dev'
-Requires-Dist: strawberry-graphql[debug-server,opentelemetry]==0.267.0; extra == 'dev'
+Requires-Dist: strawberry-graphql[debug-server,opentelemetry]==0.269.0; extra == 'dev'
 Requires-Dist: tabulate; extra == 'dev'
 Requires-Dist: tox-uv==1.11.3; extra == 'dev'
 Requires-Dist: tox==4.18.1; extra == 'dev'

{arize_phoenix-9.3.0 → arize_phoenix-9.5.0}/pyproject.toml

@@ -29,7 +29,7 @@ dependencies = [
   "starlette",
   "uvicorn",
   "psutil",
-  "strawberry-graphql==0.267.0",  # avoid this issue: https://github.com/strawberry-graphql/strawberry/issues/3854
+  "strawberry-graphql==0.269.0",  # fix relay ids
   "pyarrow",
   "typing-extensions>=4.6",
   "scipy",
@@ -88,7 +88,7 @@ dev = [
   "pytest-postgresql",
   "asyncpg",
   "psycopg[binary,pool]",
-  "strawberry-graphql[debug-server,opentelemetry]==0.267.0",  # avoid this issue: https://github.com/strawberry-graphql/strawberry/issues/3854
+  "strawberry-graphql[debug-server,opentelemetry]==0.269.0",  # fix relay ids
   "pre-commit",
   "arize[AutoEmbeddings, LLM_Evaluation]",
   "llama-index>=0.10.3",
@@ -137,7 +137,7 @@ container = [
   "opentelemetry-instrumentation-sqlalchemy",
   "opentelemetry-instrumentation-grpc",
   "py-grpc-prometheus",
-  "strawberry-graphql[opentelemetry]==0.267.0",  # avoid this issue: https://github.com/strawberry-graphql/strawberry/issues/3854
+  "strawberry-graphql[opentelemetry]==0.269.0",  # fix relay ids
   "uvloop; platform_system != 'Windows'",
   "fast-hdbscan>=0.2.0",
   "numba>=0.60.0",  # https://github.com/astral-sh/uv/issues/6281

{arize_phoenix-9.3.0 → arize_phoenix-9.5.0}/src/phoenix/db/README.md

@@ -37,7 +37,7 @@ alembic upgrade head
 
 If the above command fails, it may be necessary to undo partially applied changes from a failed migration by first running down-migrations. This can be accomplished by identifying the ID of the migration revision you wish to return to. Revisions are defined [here](./migrations/versions/).
 
-⚠️ Running down-migrations can result in lost data. Only run down-migrations if you know what you are doing and consider backing up your database first. If you have any questions or doubts, contact the Phoenix team in the `#phoenix-support` channel of the [Arize AI Slack community](https://join.slack.com/t/arize-ai/shared_invite/zt-1px8dcmlf-fmThhDFD_V_48oU7ALan4Q) or via GitHub.
+⚠️ Running down-migrations can result in lost data. Only run down-migrations if you know what you are doing and consider backing up your database first. If you have any questions or doubts, contact the Phoenix team in the `#phoenix-support` channel of the [Arize AI Slack community](https://arize-ai.slack.com/join/shared_invite/zt-2w57bhem8-hq24MB6u7yE_ZF_ilOYSBw#/shared-invite/email) or via GitHub.
 
 ```bash
 alembic downgrade <revision-id>

{arize_phoenix-9.3.0 → arize_phoenix-9.5.0}/src/phoenix/experiments/functions.py

@@ -116,8 +116,17 @@ def run_experiment(
     - `metadata`: Metadata associated with the dataset example
     - `example`: The dataset `Example` object with all associated fields
 
-    An `evaluator` is either a synchronous or asynchronous function that returns either a boolean
-    or numeric "score". If the `evaluator` is a function of one argument then that argument will be
+    An `evaluator` is either a synchronous or asynchronous function that returns an evaluation
+    result object, which can take any of the following forms.
+
+    - phoenix.experiments.types.EvaluationResult with optional fields for score, label, explanation
+      and metadata
+    - a `bool`, which will be interpreted as a score of 0 or 1 plus a label of "True" or "False"
+    - a `float`, which will be interpreted as a score
+    - a `str`, which will be interpreted as a label
+    - a 2-`tuple` of (`float`, `str`), which will be interpreted as (score, explanation)
+
+    If the `evaluator` is a function of one argument then that argument will be
     bound to the `output` of the task. Alternatively, the `evaluator` can be a function of any
     combination of specific argument names that will be bound to special values:
 

{arize_phoenix-9.3.0 → arize_phoenix-9.5.0}/src/phoenix/server/api/types/Project.py

@@ -1,13 +1,14 @@
 from __future__ import annotations
 
 import operator
-from datetime import datetime
+from datetime import datetime, timedelta
 from typing import TYPE_CHECKING, Annotated, Any, ClassVar, Optional
 
 import strawberry
 from aioitertools.itertools import islice
 from openinference.semconv.trace import SpanAttributes
 from sqlalchemy import desc, distinct, func, or_, select
+from sqlalchemy.dialects import postgresql, sqlite
 from sqlalchemy.sql.elements import ColumnElement
 from sqlalchemy.sql.expression import tuple_
 from strawberry import ID, UNSET, Private, lazy
@@ -17,6 +18,7 @@ from typing_extensions import assert_never
 
 from phoenix.datetime_utils import right_open_time_range
 from phoenix.db import models
+from phoenix.db.helpers import SupportedSQLDialect
 from phoenix.server.api.context import Context
 from phoenix.server.api.input_types.ProjectSessionSort import (
     ProjectSessionColumn,
@@ -38,6 +40,7 @@ from phoenix.server.api.types.pagination import (
 from phoenix.server.api.types.ProjectSession import ProjectSession, to_gql_project_session
 from phoenix.server.api.types.SortDir import SortDir
 from phoenix.server.api.types.Span import Span
+from phoenix.server.api.types.TimeSeries import TimeSeries, TimeSeriesDataPoint
 from phoenix.server.api.types.Trace import Trace
 from phoenix.server.api.types.ValidationResult import ValidationResult
 from phoenix.trace.dsl import SpanFilter
@@ -528,19 +531,46 @@ class Project(Node):
         return info.context.last_updated_at.get(self._table, self.project_rowid)
 
     @strawberry.field
-    async def validate_span_filter_condition(self, condition: str) -> ValidationResult:
+    async def validate_span_filter_condition(
+        self,
+        info: Info[Context, None],
+        condition: str,
+    ) -> ValidationResult:
+        """Validates a span filter condition by attempting to compile it for both SQLite and PostgreSQL.
+
+        This method checks if the provided filter condition is syntactically valid and can be compiled
+        into SQL queries for both SQLite and PostgreSQL databases. It does not execute the query,
+        only validates its syntax. Any exception during compilation (syntax errors, invalid expressions,
+        etc.) will result in an invalid validation result.
+
+        Args:
+            condition (str): The span filter condition string to validate.
+
+        Returns:
+            ValidationResult: A result object containing:
+                - is_valid (bool): True if the condition is valid, False otherwise
+                - error_message (Optional[str]): Error message if validation fails, None if valid
+        """  # noqa: E501
         # This query is too expensive to run on every validation
         # valid_eval_names = await self.span_annotation_names()
         try:
-            SpanFilter(
+            span_filter = SpanFilter(
                 condition=condition,
                 # valid_eval_names=valid_eval_names,
             )
+            stmt = span_filter(select(models.Span))
+            dialect = info.context.db.dialect
+            if dialect is SupportedSQLDialect.POSTGRESQL:
+                str(stmt.compile(dialect=sqlite.dialect()))  # type: ignore[no-untyped-call]
+            elif dialect is SupportedSQLDialect.SQLITE:
+                str(stmt.compile(dialect=postgresql.dialect()))  # type: ignore[no-untyped-call]
+            else:
+                assert_never(dialect)
             return ValidationResult(is_valid=True, error_message=None)
-        except SyntaxError as e:
+        except Exception as e:
             return ValidationResult(
                 is_valid=False,
-                error_message=e.msg,
+                error_message=str(e),
             )
 
     @strawberry.field
@@ -610,6 +640,97 @@ class Project(Node):
             )
         return updated_at
 
+    @strawberry.field(
+        description="Hourly span count for the project.",
+    )  # type: ignore
+    async def span_count_time_series(
+        self,
+        info: Info[Context, None],
+        time_range: Optional[TimeRange] = UNSET,
+    ) -> SpanCountTimeSeries:
+        """Returns a time series of span counts grouped by hour for the project.
+
+        This field provides hourly aggregated span counts, which can be useful for
+        visualizing span activity over time. The data points represent the number
+        of spans that started in each hour.
+
+        Args:
+            info: The GraphQL info object containing context information.
+            time_range: Optional time range to filter the spans. If provided, only
+                spans that started within this range will be counted.
+
+        Returns:
+            A SpanCountTimeSeries object containing data points with timestamps
+            (rounded to the nearest hour) and corresponding span counts.
+
+        Notes:
+            - The timestamps are rounded down to the nearest hour.
+            - If a time range is provided, the start time is rounded down to the
+              nearest hour, and the end time is rounded up to the nearest hour.
+            - The SQL query is optimized for both PostgreSQL and SQLite databases.
+        """
+        # Determine the appropriate SQL function to truncate timestamps to hours
+        # based on the database dialect
+        if info.context.db.dialect is SupportedSQLDialect.POSTGRESQL:
+            # PostgreSQL uses date_trunc for timestamp truncation
+            hour = func.date_trunc("hour", models.Span.start_time)
+        elif info.context.db.dialect is SupportedSQLDialect.SQLITE:
+            # SQLite uses strftime for timestamp formatting
+            hour = func.strftime("%Y-%m-%dT%H:00:00.000+00:00", models.Span.start_time)
+        else:
+            assert_never(info.context.db.dialect)
+
+        # Build the base query to count spans grouped by hour
+        stmt = (
+            select(hour, func.count())
+            .join(models.Trace)
+            .where(models.Trace.project_rowid == self.project_rowid)
+            .group_by(hour)
+            .order_by(hour)
+        )
+
+        # Apply time range filtering if provided
+        if time_range:
+            if t := time_range.start:
+                # Round down to nearest hour for the start time
+                start = t.replace(minute=0, second=0, microsecond=0)
+                stmt = stmt.where(start <= models.Span.start_time)
+            if t := time_range.end:
+                # Round up to nearest hour for the end time
+                # If the time is already at the start of an hour, use it as is
+                if t.minute == 0 and t.second == 0 and t.microsecond == 0:
+                    end = t
+                else:
+                    # Otherwise, round up to the next hour
+                    end = t.replace(minute=0, second=0, microsecond=0) + timedelta(hours=1)
+                stmt = stmt.where(models.Span.start_time < end)
+
+        # Execute the query and convert the results to a time series
+        async with info.context.db() as session:
+            data = await session.stream(stmt)
+            return SpanCountTimeSeries(
+                data=[
+                    TimeSeriesDataPoint(
+                        timestamp=_as_datetime(t),
+                        value=v,
+                    )
+                    async for t, v in data
+                ]
+            )
+
+
+@strawberry.type
+class SpanCountTimeSeries(TimeSeries):
+    """A time series of span count"""
+
 
 INPUT_VALUE = SpanAttributes.INPUT_VALUE.split(".")
 OUTPUT_VALUE = SpanAttributes.OUTPUT_VALUE.split(".")
+
+
+def _as_datetime(value: Any) -> datetime:
+    if isinstance(value, datetime):
+        return value
+    if isinstance(value, str):
+        return datetime.fromisoformat(value)
+    raise ValueError(f"Cannot convert {value} to datetime")

{arize_phoenix-9.3.0 → arize_phoenix-9.5.0}/src/phoenix/server/main.py

@@ -92,7 +92,7 @@ _WELCOME_MESSAGE = Environment(loader=BaseLoader()).from_string("""
 
 |
 |  🌎 Join our Community 🌎
-|  https://join.slack.com/t/arize-ai/shared_invite/zt-1px8dcmlf-fmThhDFD_V_48oU7ALan4Q
+|  https://arize-ai.slack.com/join/shared_invite/zt-2w57bhem8-hq24MB6u7yE_ZF_ilOYSBw#/shared-invite/email
 |
 |  ⭐️ Leave us a Star ⭐️
 |  https://github.com/Arize-ai/phoenix

arize_phoenix-9.5.0/src/phoenix/server/static/.vite/manifest.json

@@ -0,0 +1,87 @@
+{
+  "_components-DpK7N6zE.js": {
+    "file": "assets/components-DpK7N6zE.js",
+    "name": "components",
+    "imports": [
+      "_vendor-CToBXdDM.js",
+      "_pages-jHwPRLA2.js",
+      "_vendor-arizeai-BhbMHqQs.js",
+      "_vendor-codemirror-CeLHFooz.js",
+      "_vendor-three-C5WAXd5r.js"
+    ]
+  },
+  "_pages-jHwPRLA2.js": {
+    "file": "assets/pages-jHwPRLA2.js",
+    "name": "pages",
+    "imports": [
+      "_vendor-CToBXdDM.js",
+      "_vendor-arizeai-BhbMHqQs.js",
+      "_components-DpK7N6zE.js",
+      "_vendor-codemirror-CeLHFooz.js",
+      "_vendor-recharts-PlWJHgM9.js"
+    ]
+  },
+  "_vendor-CToBXdDM.js": {
+    "file": "assets/vendor-CToBXdDM.js",
+    "name": "vendor",
+    "imports": [
+      "_vendor-three-C5WAXd5r.js"
+    ],
+    "css": [
+      "assets/vendor-WIZid84E.css"
+    ]
+  },
+  "_vendor-WIZid84E.css": {
+    "file": "assets/vendor-WIZid84E.css",
+    "src": "_vendor-WIZid84E.css"
+  },
+  "_vendor-arizeai-BhbMHqQs.js": {
+    "file": "assets/vendor-arizeai-BhbMHqQs.js",
+    "name": "vendor-arizeai",
+    "imports": [
+      "_vendor-CToBXdDM.js"
+    ]
+  },
+  "_vendor-codemirror-CeLHFooz.js": {
+    "file": "assets/vendor-codemirror-CeLHFooz.js",
+    "name": "vendor-codemirror",
+    "imports": [
+      "_vendor-CToBXdDM.js",
+      "_vendor-shiki-CPwL2jwA.js"
+    ]
+  },
+  "_vendor-recharts-PlWJHgM9.js": {
+    "file": "assets/vendor-recharts-PlWJHgM9.js",
+    "name": "vendor-recharts",
+    "imports": [
+      "_vendor-CToBXdDM.js"
+    ]
+  },
+  "_vendor-shiki-CPwL2jwA.js": {
+    "file": "assets/vendor-shiki-CPwL2jwA.js",
+    "name": "vendor-shiki",
+    "imports": [
+      "_vendor-CToBXdDM.js"
+    ]
+  },
+  "_vendor-three-C5WAXd5r.js": {
+    "file": "assets/vendor-three-C5WAXd5r.js",
+    "name": "vendor-three"
+  },
+  "index.tsx": {
+    "file": "assets/index-BXA0RjaV.js",
+    "name": "index",
+    "src": "index.tsx",
+    "isEntry": true,
+    "imports": [
+      "_vendor-CToBXdDM.js",
+      "_vendor-arizeai-BhbMHqQs.js",
+      "_pages-jHwPRLA2.js",
+      "_components-DpK7N6zE.js",
+      "_vendor-three-C5WAXd5r.js",
+      "_vendor-codemirror-CeLHFooz.js",
+      "_vendor-shiki-CPwL2jwA.js",
+      "_vendor-recharts-PlWJHgM9.js"
+    ]
+  }
+}