arize-phoenix 11.10.1__tar.gz → 11.12.0__tar.gz

{arize_phoenix-11.10.1 → arize_phoenix-11.12.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: arize-phoenix
-Version: 11.10.1
+Version: 11.12.0
 Summary: AI Observability and Evaluation
 Project-URL: Documentation, https://arize.com/docs/phoenix/
 Project-URL: Issues, https://github.com/Arize-ai/phoenix/issues

arize_phoenix-11.12.0/src/phoenix/server/api/input_types/PromptFilter.py

@@ -0,0 +1,14 @@
+from enum import Enum
+
+import strawberry
+
+
+@strawberry.enum
+class PromptFilterColumn(Enum):
+    name = "name"
+
+
+@strawberry.input(description="The filter key and value for prompt connections")
+class PromptFilter:
+    col: PromptFilterColumn
+    value: str

{arize_phoenix-11.10.1 → arize_phoenix-11.12.0}/src/phoenix/server/api/mutations/user_mutations.py

@@ -153,7 +153,7 @@ class UserMutationMixin:
                 logger.error(f"Failed to send welcome email: {error}")
         return UserMutationPayload(user=to_gql_user(user))
 
-    @strawberry.mutation(permission_classes=[IsNotReadOnly, IsAdmin, IsLocked])  # type: ignore
+    @strawberry.mutation(permission_classes=[IsNotReadOnly, IsAdmin])  # type: ignore
     async def patch_user(
         self,
         info: Info[Context, None],
@@ -199,7 +199,7 @@ class UserMutationMixin:
             await info.context.log_out(user.id)
         return UserMutationPayload(user=to_gql_user(user))
 
-    @strawberry.mutation(permission_classes=[IsNotReadOnly, IsLocked])  # type: ignore
+    @strawberry.mutation(permission_classes=[IsNotReadOnly])  # type: ignore
     async def patch_viewer(
         self,
         info: Info[Context, None],

{arize_phoenix-11.10.1 → arize_phoenix-11.12.0}/src/phoenix/server/api/queries.py

@@ -1,12 +1,13 @@
 import re
 from collections import defaultdict
 from datetime import datetime
-from typing import Iterable, Iterator, Optional, Union, cast
+from typing import Iterable, Iterator, Optional, Union
+from typing import cast as type_cast
 
 import numpy as np
 import numpy.typing as npt
 import strawberry
-from sqlalchemy import and_, distinct, func, select, text
+from sqlalchemy import String, and_, cast, distinct, func, select, text
 from sqlalchemy.orm import joinedload
 from starlette.authentication import UnauthenticatedUser
 from strawberry import ID, UNSET
@@ -41,6 +42,7 @@ from phoenix.server.api.input_types.DatasetSort import DatasetSort
 from phoenix.server.api.input_types.InvocationParameters import InvocationParameter
 from phoenix.server.api.input_types.ProjectFilter import ProjectFilter
 from phoenix.server.api.input_types.ProjectSort import ProjectColumn, ProjectSort
+from phoenix.server.api.input_types.PromptFilter import PromptFilter
 from phoenix.server.api.types.AnnotationConfig import AnnotationConfig, to_gql_annotation_config
 from phoenix.server.api.types.Cluster import Cluster, to_gql_clusters
 from phoenix.server.api.types.Dataset import Dataset, to_gql_dataset
@@ -77,6 +79,7 @@ from phoenix.server.api.types.Prompt import Prompt, to_gql_prompt_from_orm
 from phoenix.server.api.types.PromptLabel import PromptLabel, to_gql_prompt_label
 from phoenix.server.api.types.PromptVersion import PromptVersion, to_gql_prompt_version
 from phoenix.server.api.types.PromptVersionTag import PromptVersionTag, to_gql_prompt_version_tag
+from phoenix.server.api.types.ServerStatus import ServerStatus
 from phoenix.server.api.types.SortDir import SortDir
 from phoenix.server.api.types.Span import Span
 from phoenix.server.api.types.SpanAnnotation import SpanAnnotation, to_gql_span_annotation
@@ -728,6 +731,7 @@ class Query:
         last: Optional[int] = UNSET,
         after: Optional[CursorString] = UNSET,
         before: Optional[CursorString] = UNSET,
+        filter: Optional[PromptFilter] = UNSET,
     ) -> Connection[Prompt]:
         args = ConnectionArgs(
             first=first,
@@ -736,6 +740,14 @@ class Query:
             before=before if isinstance(before, CursorString) else None,
         )
         stmt = select(models.Prompt)
+        if filter:
+            column = getattr(models.Prompt, filter.col.value)
+            # Cast Identifier columns to String for ilike operations
+            if filter.col.value == "name":
+                column = cast(column, String)
+            stmt = stmt.where(column.ilike(f"%{filter.value}%")).order_by(
+                models.Prompt.updated_at.desc()
+            )
         async with info.context.db() as session:
             orm_prompts = await session.stream_scalars(stmt)
             data = [to_gql_prompt_from_orm(orm_prompt) async for orm_prompt in orm_prompts]
@@ -994,7 +1006,7 @@ class Query:
             """).bindparams(nspname=nspname)
             try:
                 async with info.context.db() as session:
-                    stats = cast(Iterable[tuple[str, int]], await session.execute(stmt))
+                    stats = type_cast(Iterable[tuple[str, int]], await session.execute(stmt))
             except Exception:
                 # TODO: temporary workaround until we can reproduce the error
                 return []
@@ -1005,6 +1017,15 @@ class Query:
             for table_name, num_bytes in stats
         ]
 
+    @strawberry.field
+    async def server_status(
+        self,
+        info: Info[Context, None],
+    ) -> ServerStatus:
+        return ServerStatus(
+            insufficient_storage=info.context.db.should_not_insert_or_update,
+        )
+
     @strawberry.field
     def validate_regular_expression(self, regex: str) -> ValidationResult:
         try:

{arize_phoenix-11.10.1 → arize_phoenix-11.12.0}/src/phoenix/server/api/types/GenerativeModel.py

@@ -6,14 +6,19 @@ import strawberry
 from openinference.semconv.trace import OpenInferenceLLMProviderValues
 from sqlalchemy import inspect
 from strawberry.relay import Node, NodeID
+from strawberry.relay.types import GlobalID
 from strawberry.types import Info
-from typing_extensions import assert_never
+from strawberry.types.unset import UNSET
+from typing_extensions import TypeAlias, assert_never
 
 from phoenix.db import models
 from phoenix.server.api.context import Context
+from phoenix.server.api.exceptions import BadRequest
+from phoenix.server.api.input_types.TimeRange import TimeRange
 from phoenix.server.api.types.CostBreakdown import CostBreakdown
 from phoenix.server.api.types.GenerativeProvider import GenerativeProviderKey
 from phoenix.server.api.types.ModelInterface import ModelInterface
+from phoenix.server.api.types.node import from_global_id
 from phoenix.server.api.types.SpanCostDetailSummaryEntry import SpanCostDetailSummaryEntry
 from phoenix.server.api.types.SpanCostSummary import SpanCostSummary
 from phoenix.server.api.types.TokenPrice import TokenKind, TokenPrice
@@ -25,6 +30,11 @@ class GenerativeModelKind(Enum):
     BUILT_IN = "BUILT_IN"
 
 
+ProjectId: TypeAlias = int
+TimeRangeKey: TypeAlias = tuple[Optional[datetime], Optional[datetime]]
+CachedCostSummaryKey: TypeAlias = tuple[Optional[ProjectId], TimeRangeKey]
+
+
 @strawberry.type
 class GenerativeModel(Node, ModelInterface):
     id_attr: NodeID[int]
@@ -37,6 +47,18 @@ class GenerativeModel(Node, ModelInterface):
     provider_key: Optional[GenerativeProviderKey]
     costs: strawberry.Private[Optional[list[models.TokenPrice]]] = None
     start_time: Optional[datetime] = None
+    cached_cost_summary: strawberry.Private[
+        Optional[dict[CachedCostSummaryKey, SpanCostSummary]]
+    ] = None
+
+    def add_cached_cost_summary(
+        self, project_id: Optional[int], time_range: TimeRange, cost_summary: SpanCostSummary
+    ) -> None:
+        if self.cached_cost_summary is None:
+            self.cached_cost_summary = {}
+        time_range_key = (time_range.start, time_range.end) if time_range else (None, None)
+        cache_key = (project_id, time_range_key)
+        self.cached_cost_summary[cache_key] = cost_summary
 
     @strawberry.field
     async def token_prices(self) -> list[TokenPrice]:
@@ -55,7 +77,28 @@ class GenerativeModel(Node, ModelInterface):
         return token_prices
 
     @strawberry.field
-    async def cost_summary(self, info: Info[Context, None]) -> SpanCostSummary:
+    async def cost_summary(
+        self,
+        info: Info[Context, None],
+        project_id: Optional[GlobalID] = UNSET,
+        time_range: Optional[TimeRange] = UNSET,
+    ) -> SpanCostSummary:
+        if self.cached_cost_summary is not None:
+            time_range_key = (time_range.start, time_range.end) if time_range else (None, None)
+            project_rowid: Optional[int] = None
+            if project_id:
+                type_name, project_rowid = from_global_id(project_id)
+                if type_name != models.Project.__name__:
+                    raise BadRequest("Invalid Project ID")
+            cache_key = (project_rowid, time_range_key)
+            if cache_key in self.cached_cost_summary:
+                return self.cached_cost_summary[cache_key]
+
+        if time_range or project_id:
+            raise BadRequest(
+                "Cost summaries for specific projects or time ranges are not yet implemented"
+            )
+
         loader = info.context.data_loaders.span_cost_summary_by_generative_model
         summary = await loader.load(self.id_attr)
         return SpanCostSummary(
@@ -98,7 +141,9 @@ class GenerativeModel(Node, ModelInterface):
         return await info.context.data_loaders.last_used_times_by_generative_model_id.load(model_id)
 
 
-def to_gql_generative_model(model: models.GenerativeModel) -> GenerativeModel:
+def to_gql_generative_model(
+    model: models.GenerativeModel,
+) -> GenerativeModel:
     costs_are_loaded = isinstance(inspect(model).attrs.token_prices.loaded_value, list)
     name_pattern = model.name_pattern.pattern
     assert isinstance(name_pattern, str)

{arize_phoenix-11.10.1 → arize_phoenix-11.12.0}/src/phoenix/server/api/types/Project.py

@@ -33,6 +33,7 @@ from phoenix.server.api.types.AnnotationConfig import AnnotationConfig, to_gql_a
 from phoenix.server.api.types.AnnotationSummary import AnnotationSummary
 from phoenix.server.api.types.CostBreakdown import CostBreakdown
 from phoenix.server.api.types.DocumentEvaluationSummary import DocumentEvaluationSummary
+from phoenix.server.api.types.GenerativeModel import GenerativeModel, to_gql_generative_model
 from phoenix.server.api.types.pagination import (
     ConnectionArgs,
     Cursor,
@@ -1326,6 +1327,130 @@ class Project(Node):
             names=sorted(list(unique_names)),
         )
 
+    @strawberry.field
+    async def top_models_by_cost(
+        self,
+        info: Info[Context, None],
+        time_range: TimeRange,
+    ) -> list[GenerativeModel]:
+        if time_range.start is None:
+            raise BadRequest("Start time is required")
+
+        async with info.context.db() as session:
+            stmt = (
+                select(
+                    models.GenerativeModel,
+                    func.sum(models.SpanCost.total_tokens).label("total_tokens"),
+                    func.sum(models.SpanCost.prompt_tokens).label("prompt_tokens"),
+                    func.sum(models.SpanCost.completion_tokens).label("completion_tokens"),
+                    func.sum(models.SpanCost.total_cost).label("total_cost"),
+                    func.sum(models.SpanCost.prompt_cost).label("prompt_cost"),
+                    func.sum(models.SpanCost.completion_cost).label("completion_cost"),
+                )
+                .join(
+                    models.SpanCost,
+                    models.SpanCost.model_id == models.GenerativeModel.id,
+                )
+                .join(
+                    models.Trace,
+                    models.SpanCost.trace_rowid == models.Trace.id,
+                )
+                .where(models.Trace.project_rowid == self.project_rowid)
+                .where(models.SpanCost.model_id.isnot(None))
+                .where(models.SpanCost.span_start_time >= time_range.start)
+                .group_by(models.SpanCost.model_id)
+                .order_by(func.sum(models.SpanCost.total_cost).desc())
+            )
+            if time_range.end:
+                stmt = stmt.where(models.SpanCost.span_start_time < time_range.end)
+            results: list[GenerativeModel] = []
+            async for (
+                model,
+                total_tokens,
+                prompt_tokens,
+                completion_tokens,
+                total_cost,
+                prompt_cost,
+                completion_cost,
+            ) in await session.stream(stmt):
+                cost_summary = SpanCostSummary(
+                    prompt=CostBreakdown(tokens=prompt_tokens, cost=prompt_cost),
+                    completion=CostBreakdown(tokens=completion_tokens, cost=completion_cost),
+                    total=CostBreakdown(tokens=total_tokens, cost=total_cost),
+                )
+                cache_time_range = TimeRange(
+                    start=time_range.start,
+                    end=time_range.end,
+                )
+                gql_model = to_gql_generative_model(model)
+                gql_model.add_cached_cost_summary(
+                    self.project_rowid, cache_time_range, cost_summary
+                )
+                results.append(gql_model)
+            return results
+
+    @strawberry.field
+    async def top_models_by_token_count(
+        self,
+        info: Info[Context, None],
+        time_range: TimeRange,
+    ) -> list[GenerativeModel]:
+        if time_range.start is None:
+            raise BadRequest("Start time is required")
+
+        async with info.context.db() as session:
+            stmt = (
+                select(
+                    models.GenerativeModel,
+                    func.sum(models.SpanCost.total_tokens).label("total_tokens"),
+                    func.sum(models.SpanCost.prompt_tokens).label("prompt_tokens"),
+                    func.sum(models.SpanCost.completion_tokens).label("completion_tokens"),
+                    func.sum(models.SpanCost.total_cost).label("total_cost"),
+                    func.sum(models.SpanCost.prompt_cost).label("prompt_cost"),
+                    func.sum(models.SpanCost.completion_cost).label("completion_cost"),
+                )
+                .join(
+                    models.SpanCost,
+                    models.SpanCost.model_id == models.GenerativeModel.id,
+                )
+                .join(
+                    models.Trace,
+                    models.SpanCost.trace_rowid == models.Trace.id,
+                )
+                .where(models.Trace.project_rowid == self.project_rowid)
+                .where(models.SpanCost.model_id.isnot(None))
+                .where(models.SpanCost.span_start_time >= time_range.start)
+                .group_by(models.SpanCost.model_id)
+                .order_by(func.sum(models.SpanCost.total_tokens).desc())
+            )
+            if time_range.end:
+                stmt = stmt.where(models.SpanCost.span_start_time < time_range.end)
+            results: list[GenerativeModel] = []
+            async for (
+                model,
+                total_tokens,
+                prompt_tokens,
+                completion_tokens,
+                total_cost,
+                prompt_cost,
+                completion_cost,
+            ) in await session.stream(stmt):
+                cost_summary = SpanCostSummary(
+                    prompt=CostBreakdown(tokens=prompt_tokens, cost=prompt_cost),
+                    completion=CostBreakdown(tokens=completion_tokens, cost=completion_cost),
+                    total=CostBreakdown(tokens=total_tokens, cost=total_cost),
+                )
+                cache_time_range = TimeRange(
+                    start=time_range.start,
+                    end=time_range.end,
+                )
+                gql_model = to_gql_generative_model(model)
+                gql_model.add_cached_cost_summary(
+                    self.project_rowid, cache_time_range, cost_summary
+                )
+                results.append(gql_model)
+            return results
+
 
 @strawberry.type
 class SpanCountTimeSeriesDataPoint:

arize_phoenix-11.12.0/src/phoenix/server/api/types/ServerStatus.py

@@ -0,0 +1,6 @@
+import strawberry
+
+
+@strawberry.type
+class ServerStatus:
+    insufficient_storage: bool

{arize_phoenix-11.10.1 → arize_phoenix-11.12.0}/src/phoenix/server/app.py

@@ -56,6 +56,8 @@ from phoenix.config import (
     SERVER_DIR,
     OAuth2ClientConfig,
     get_env_csrf_trusted_origins,
+    get_env_database_allocated_storage_capacity_gibibytes,
+    get_env_database_usage_insertion_blocking_threshold_percentage,
     get_env_fastapi_middleware_paths,
     get_env_gql_extension_paths,
     get_env_grpc_interceptor_paths,
@@ -239,6 +241,10 @@ class AppConfig(NamedTuple):
     """ FullStory organization ID for web analytics tracking """
     management_url: Optional[str] = None
     """ URL for a phoenix management interface, only visible to management users """
+    support_email: Optional[str] = None
+    """ Support email address for user assistance """
+    has_db_threshold: bool = False
+    """ Whether the database has a threshold for usage """
 
 
 class Static(StaticFiles):
@@ -306,6 +312,8 @@ class Static(StaticFiles):
                     "auto_login_idp_name": self._app_config.auto_login_idp_name,
                     "fullstory_org": self._app_config.fullstory_org,
                     "management_url": self._app_config.management_url,
+                    "support_email": self._app_config.support_email,
+                    "has_db_threshold": self._app_config.has_db_threshold,
                 },
             )
         except Exception as e:
@@ -1072,6 +1080,11 @@ def create_app(
                     auto_login_idp_name=auto_login_idp_name,
                     fullstory_org=Settings.fullstory_org,
                     management_url=management_url,
+                    support_email=get_env_support_email(),
+                    has_db_threshold=bool(
+                        get_env_database_allocated_storage_capacity_gibibytes()
+                        and get_env_database_usage_insertion_blocking_threshold_percentage()
+                    ),
                 ),
             ),
             name="static",

arize_phoenix-11.12.0/src/phoenix/server/static/.vite/manifest.json

@@ -0,0 +1,87 @@
+{
+  "_components-E49JuzFB.js": {
+    "file": "assets/components-E49JuzFB.js",
+    "name": "components",
+    "imports": [
+      "_vendor-_6rG8OMg.js",
+      "_pages-C-8lZ5cv.js",
+      "_vendor-arizeai-BznCmJFh.js",
+      "_vendor-codemirror-29fWLPAy.js",
+      "_vendor-three-C5WAXd5r.js"
+    ]
+  },
+  "_pages-C-8lZ5cv.js": {
+    "file": "assets/pages-C-8lZ5cv.js",
+    "name": "pages",
+    "imports": [
+      "_vendor-_6rG8OMg.js",
+      "_vendor-arizeai-BznCmJFh.js",
+      "_components-E49JuzFB.js",
+      "_vendor-codemirror-29fWLPAy.js",
+      "_vendor-recharts-Cu431IpB.js"
+    ]
+  },
+  "_vendor-CqDb5u4o.css": {
+    "file": "assets/vendor-CqDb5u4o.css",
+    "src": "_vendor-CqDb5u4o.css"
+  },
+  "_vendor-_6rG8OMg.js": {
+    "file": "assets/vendor-_6rG8OMg.js",
+    "name": "vendor",
+    "imports": [
+      "_vendor-three-C5WAXd5r.js"
+    ],
+    "css": [
+      "assets/vendor-CqDb5u4o.css"
+    ]
+  },
+  "_vendor-arizeai-BznCmJFh.js": {
+    "file": "assets/vendor-arizeai-BznCmJFh.js",
+    "name": "vendor-arizeai",
+    "imports": [
+      "_vendor-_6rG8OMg.js"
+    ]
+  },
+  "_vendor-codemirror-29fWLPAy.js": {
+    "file": "assets/vendor-codemirror-29fWLPAy.js",
+    "name": "vendor-codemirror",
+    "imports": [
+      "_vendor-_6rG8OMg.js",
+      "_vendor-shiki-Ce9e01lU.js"
+    ]
+  },
+  "_vendor-recharts-Cu431IpB.js": {
+    "file": "assets/vendor-recharts-Cu431IpB.js",
+    "name": "vendor-recharts",
+    "imports": [
+      "_vendor-_6rG8OMg.js"
+    ]
+  },
+  "_vendor-shiki-Ce9e01lU.js": {
+    "file": "assets/vendor-shiki-Ce9e01lU.js",
+    "name": "vendor-shiki",
+    "imports": [
+      "_vendor-_6rG8OMg.js"
+    ]
+  },
+  "_vendor-three-C5WAXd5r.js": {
+    "file": "assets/vendor-three-C5WAXd5r.js",
+    "name": "vendor-three"
+  },
+  "index.tsx": {
+    "file": "assets/index-CUhYDnlT.js",
+    "name": "index",
+    "src": "index.tsx",
+    "isEntry": true,
+    "imports": [
+      "_vendor-_6rG8OMg.js",
+      "_vendor-arizeai-BznCmJFh.js",
+      "_pages-C-8lZ5cv.js",
+      "_components-E49JuzFB.js",
+      "_vendor-three-C5WAXd5r.js",
+      "_vendor-codemirror-29fWLPAy.js",
+      "_vendor-shiki-Ce9e01lU.js",
+      "_vendor-recharts-Cu431IpB.js"
+    ]
+  }
+}