agi-med-common 5.0.15__tar.gz → 5.0.17__tar.gz

{agi_med_common-5.0.15 → agi_med_common-5.0.17}/PKG-INFO

@@ -1,12 +1,11 @@
 Metadata-Version: 2.4
 Name: agi_med_common
-Version: 5.0.15
+Version: 5.0.17
 Summary: Сommon for agi-med team
 Author: AGI-MED-TEAM
 Requires-Python: >=3.11
 Description-Content-Type: text/markdown
 Requires-Dist: pydantic<=3.0.0,>=2.9.2
-Requires-Dist: loguru~=0.7.2
 Requires-Dist: tqdm==4.67.*
 
 # agi-med-common

{agi_med_common-5.0.15 → agi_med_common-5.0.17}/pyproject.toml

@@ -5,17 +5,17 @@ build-backend = "setuptools.build_meta"
 [project]
 name = 'agi_med_common'
 requires-python = ">= 3.11"
-
 authors = [{ name = 'AGI-MED-TEAM' }]
 description = 'Сommon for agi-med team'
 readme = 'README.md'
-
-dynamic = ['version', 'dependencies']
+dynamic = ['version']
+dependencies = [
+    "pydantic>=2.9.2,<=3.0.0",
+    "tqdm==4.67.*",
+]
 
 [tool.setuptools.packages.find]
 where = ["src"]
 
-
 [tool.setuptools.dynamic]
 version = { attr = "agi_med_common.__version__" }
-dependencies = { file = ['requirements.txt'] }

{agi_med_common-5.0.15 → agi_med_common-5.0.17}/src/agi_med_common/__init__.py

@@ -1,12 +1,17 @@
-__version__ = "5.0.15"
+__version__ = "5.0.17"
 
-from .logger import LogLevelEnum, logger_init, log_llm_error
 from .models import (
     MTRSLabelEnum,
     ChatItem,
     InnerContextItem,
     OuterContextItem,
     ReplicaItem,
+    Chat,
+    Context,
+    ChatMessage,
+    AIMessage,
+    HumanMessage,
+    MiscMessage,
 )
 from .models.widget import Widget
 from .file_storage import FileStorage, ResourceId
@@ -19,5 +24,5 @@ from .models.tracks import TrackInfo, DomainInfo
 from .api.chat_manager_api import ChatManagerAPI
 from .api.content_interpreter_api import ContentInterpreterAPI, Interpretation
 from .api.content_interpreter_remote_api import ContentInterpreterRemoteAPI
-from .api.text_processor_api import TextProcessorAPI
-from .api.critic_api import CriticAPI
+from .api.text_generator_api import TextGeneratorAPI
+from .api.classifier_api import ClassifierAPI

agi_med_common-5.0.17/src/agi_med_common/api/classifier_api.py

@@ -0,0 +1,14 @@
+from typing import List
+
+from agi_med_common.models import ChatItem
+
+
+Value = str
+
+
+class ClassifierAPI:
+    def get_values(self) -> List[Value]:
+        raise NotImplementedError
+
+    def evaluate(self, chat: ChatItem, request_id: str = "") -> Value:
+        raise NotImplementedError

agi_med_common-5.0.17/src/agi_med_common/api/text_generator_api.py

@@ -0,0 +1,6 @@
+from agi_med_common.models import ChatItem
+
+
+class TextGeneratorAPI:
+    def process(self, chat: ChatItem, request_id: str = "") -> str:
+        raise NotImplementedError

{agi_med_common-5.0.15 → agi_med_common-5.0.17}/src/agi_med_common/models/__init__.py

@@ -2,5 +2,6 @@ from ._base import _Base
 
 from .enums import MTRSLabelEnum, DiagnosticsXMLTagEnum, MTRSXMLTagEnum, DoctorChoiceXMLTagEnum
 from .chat_item import ChatItem, OuterContextItem, InnerContextItem, ReplicaItem
+from .chat import Chat, Context, ChatMessage, AIMessage, HumanMessage, MiscMessage
 from .base_config_models import GigaChatConfig
 from .tracks import TrackInfo, DomainInfo

agi_med_common-5.0.17/src/agi_med_common/models/chat.py

@@ -0,0 +1,267 @@
+from datetime import datetime
+from typing import Any, List, Dict, Literal
+
+from agi_med_common.models.chat_item import ChatItem, ReplicaItem, OuterContextItem
+from agi_med_common.models.widget import Widget
+from agi_med_common.type_union import TypeUnion
+from agi_med_common.utils import first_nonnull
+from pydantic import Field
+
+from ._base import _Base
+
+
+_DT_FORMAT: str = "%Y-%m-%d-%H-%M-%S"
+_EXAMPLE_DT: str = datetime(year=1970, month=1, day=1).strftime(_DT_FORMAT)
+StrDict = Dict[str, Any]
+ContentBase = str | Widget | StrDict
+Content = ContentBase | List[ContentBase]
+
+
+def now_pretty() -> str:
+    return datetime.now().strftime(_DT_FORMAT)
+
+
+class Context(_Base):
+    client_id: str = Field("", examples=["543216789"])
+    user_id: str = Field("", examples=["123456789"])
+    session_id: str = Field("", examples=["987654321"])
+    track_id: str = Field(examples=["Hello"])
+    extra: StrDict | None = Field(None, examples=[None])
+
+    def create_id(self, short: bool = False) -> str:
+        uid, sid, cid = self.user_id, self.session_id, self.client_id
+        if short:
+            return f"{cid}_{uid}_{sid}"
+        return f"client_{cid}_user_{uid}_session_{sid}"
+
+
+def _get_str_field(obj: dict, field) -> str | None:
+    if not isinstance(obj, dict):
+        return None
+    text = obj.get(field)
+    if text is not None and isinstance(text, str):
+        return text
+    return None
+
+
+def _get_text(obj: Content) -> str:
+    if isinstance(obj, str):
+        return obj
+    if isinstance(obj, list):
+        return "".join(map(_get_text, obj))
+    if isinstance(obj, dict) and obj.get("type") == "text":
+        return _get_str_field(obj, "text") or ""
+    return ""
+
+
+def _get_resource_id(obj: Content) -> str | None:
+    if isinstance(obj, list):
+        return first_nonnull(map(_get_resource_id, obj))
+    if isinstance(obj, dict) and obj.get("type") == "resource_id":
+        return _get_str_field(obj, "resource_id")
+    return None
+
+
+def _get_command(obj: Content) -> dict | None:
+    if isinstance(obj, list):
+        return first_nonnull(map(_get_command, obj))
+    if isinstance(obj, dict) and obj.get("type") == "command":
+        return _get_str_field(obj, "command")
+    return None
+
+
+def _get_widget(obj: Content) -> Widget | None:
+    if isinstance(obj, list):
+        return first_nonnull(map(_get_widget, obj))
+    if isinstance(obj, Widget):
+        return obj
+    return None
+
+
+# todo fix: generalize functions above
+
+
+class BaseMessage(_Base):
+    type: str
+    content: Content = Field("", examples=["Привет"])
+    date_time: str = Field(default_factory=now_pretty, examples=[_EXAMPLE_DT])
+    extra: StrDict | None = Field(None, examples=[None])
+
+    @property
+    def text(self) -> str:
+        return _get_text(self.content)
+
+    @property
+    def resource_id(self) -> str | None:
+        return _get_resource_id(self.content)
+
+    @property
+    def command(self) -> dict | None:
+        return _get_command(self.content)
+
+    @property
+    def widget(self) -> Widget | None:
+        return _get_widget(self.content)
+
+    @staticmethod
+    def DATETIME_FORMAT() -> str:
+        return _DT_FORMAT
+
+    def with_now_datetime(self):
+        return self.model_copy(update=dict(date_time=now_pretty()))
+
+
+class HumanMessage(BaseMessage):
+    type: Literal["human"] = "human"
+
+
+class AIMessage(BaseMessage):
+    type: Literal["ai"] = "ai"
+    state: str = Field("", examples=["COLLECTION"])
+
+
+class MiscMessage(BaseMessage):
+    type: Literal["misc"] = "misc"
+
+
+ChatMessage = TypeUnion[HumanMessage, AIMessage, MiscMessage]
+
+
+class Chat(_Base):
+    context: Context
+    messages: List[ChatMessage] = []
+
+    def create_id(self, short: bool = False) -> str:
+        return self.context.create_id(short)
+
+    @classmethod
+    def parse(self, chat_obj: str | dict) -> "Chat":
+        return _parse_chat_compat(chat_obj)
+
+
+def convert_replica_item_to_message(replica: ReplicaItem) -> ChatMessage:
+    resource_id = (replica.resource_id or None) and {"type": "resource_id", "resource_id": replica.resource_id}
+    body = replica.body
+    command = replica.command
+    widget = replica.widget
+
+    content = list(filter(None, [body, resource_id, command, widget]))
+    if len(content) == 0:
+        content = ""
+    elif len(content) == 1:
+        content = content[0]
+
+    kwargs = dict(
+        content=content,
+        date_time=replica.date_time,
+        extra=replica.extra,
+    )
+    if not replica.role:
+        return HumanMessage(**kwargs)
+    return AIMessage(
+        **kwargs,
+        state=replica.state,
+        extra=dict(
+            action=replica.action,
+            moderation=replica.moderation,
+        ),
+    )
+
+
+def convert_outer_context_to_context(octx: OuterContextItem) -> Context:
+    # legacy: eliminate
+    context = Context(
+        client_id=octx.client_id,
+        user_id=octx.user_id,
+        session_id=octx.session_id,
+        track_id=octx.track_id,
+        extra=dict(
+            sex=octx.sex,
+            age=octx.age,
+            parent_session_id=octx.parent_session_id,
+            entrypoint_key=octx.entrypoint_key,
+            language_code=octx.language_code,
+        ),
+    )
+    return context
+
+
+def convert_chat_item_to_chat(chat_item: ChatItem) -> Chat:
+    # legacy: eliminate
+    context = convert_outer_context_to_context(chat_item.outer_context)
+    messages = [convert_replica_item_to_message(replica) for replica in chat_item.inner_context.replicas]
+    res = Chat(context=context, messages=messages)
+    return res
+
+
+def convert_context_to_outer_context(context: Context) -> OuterContextItem:
+    # legacy: eliminate
+    extra = context.extra or {}
+    return OuterContextItem(
+        client_id=context.client_id,
+        user_id=context.user_id,
+        session_id=context.session_id,
+        track_id=context.track_id,
+        sex=extra.get("sex"),
+        age=extra.get("age"),
+        parent_session_id=extra.get("parent_session_id"),
+        entrypoint_key=extra.get("entrypoint_key"),
+        language_code=extra.get("language_code"),
+    )
+
+
+def convert_message_to_replica_item(message: ChatMessage) -> ReplicaItem | None:
+    # legacy: eliminate
+    m_type = message.type
+    if m_type in {"ai", "human"}:
+        role = m_type == "ai"
+    else:
+        return None
+
+    extra = message.extra or {}
+    action = extra.pop("action")
+    moderation = extra.pop("moderation")
+
+    kwargs = dict(
+        role=role,
+        body=message.body,
+        resource_id=message.resource_id,
+        command=message.command,
+        widget=message.widget,
+        date_time=message.date_time,
+        extra=extra or None,
+        state=getattr(message, "state", ""),
+        action=action,
+        moderation=moderation,
+    )
+    return ReplicaItem(**kwargs)
+
+
+def convert_chat_to_chat_item(chat: Chat) -> ChatItem:
+    # legacy: eliminate
+    return ChatItem(
+        outer_context=convert_context_to_outer_context(chat.context),
+        inner_context=dict(replicas=list(map(convert_message_to_replica_item, chat.messages))),
+    )
+
+
+def parse_chat_item_as_chat(chat_obj: str | dict) -> Chat:
+    # legacy: eliminate
+    chat_item = ChatItem.parse(chat_obj)
+    res = convert_chat_item_to_chat(chat_item)
+    return res
+
+
+def _parse_chat(chat_obj: str | dict) -> Chat:
+    if isinstance(chat_obj, dict):
+        return Chat.model_validate(chat_obj)
+
+    return Chat.model_validate_json(chat_obj)
+
+
+def _parse_chat_compat(chat_obj: str | dict) -> Chat:
+    # legacy: eliminate
+    try:
+        return _parse_chat(chat_obj)
+    except Exception:
+        return parse_chat_item_as_chat(chat_obj)

{agi_med_common-5.0.15 → agi_med_common-5.0.17}/src/agi_med_common/models/chat_item.py

@@ -1,10 +1,8 @@
-import json
 from datetime import datetime
-from typing import Any, List
+from typing import Annotated, Any, List
 
 from agi_med_common.models.widget import Widget
-from loguru import logger
-from pydantic import Field, ConfigDict, ValidationError
+from pydantic import Field, ConfigDict, BeforeValidator
 
 from ._base import _Base
 
@@ -41,6 +39,24 @@ class OuterContextItem(_Base):
     def to_dict(self) -> dict[str, Any]:
         return self.model_dump(by_alias=True)
 
+LABELS = {
+    0: "OK",
+    1: "NON_MED",
+    2: "CHILD",
+    3: "ABSURD",
+    4: "GREETING",
+    5: "RECEIPT",
+}
+
+
+def fix_deprecated_moderation(moderation):
+    if isinstance(moderation, int):
+        return LABELS.get(moderation, "OK")
+    elif isinstance(moderation, str):
+        return moderation
+    else:
+        raise ValueError(f"Unsupported moderation: {moderation} :: {type(moderation)}")
+
 
 class ReplicaItem(_Base):
     body: str = Field("", alias="Body", examples=["Привет"])
@@ -54,7 +70,9 @@ class ReplicaItem(_Base):
     state: str = Field("", alias="State", description="chat manager fsm state", examples=["COLLECTION"])
     action: str = Field("", alias="Action", description="chat manager fsm action", examples=["DIAGNOSIS"])
     # todo fix: support loading from `moderation: int`
-    moderation: str = Field("OK", alias="Moderation", description="moderation outcome", examples=["OK"])
+    moderation: Annotated[str, BeforeValidator(str)] = Field(
+        "OK", alias="Moderation", description="moderation outcome", examples=["OK"]
+    )
     extra: dict | None = Field(None, alias="Extra", examples=[None])
 
     def to_dict(self) -> dict[str, Any]:
@@ -97,38 +115,12 @@ class ChatItem(_Base):
         return [replica.to_dict().get(field, None) for replica in self.inner_context.replicas]
 
     @classmethod
-    def parse(cls, chat_text: str) -> "ChatItem":
-        return _parse_chat(chat_text)
-
-
-def _is_moderation_int_error(err):
-    # fmt: off
-    if err['type'] != 'string_type': return False
-    if err['loc'][0] != 'InnerContext': return False
-    if err['loc'][1] != 'Replicas': return False
-    if err['loc'][3] != 'Moderation': return False
-    if err['msg'] != 'Input should be a valid string': return False
-    # fmt: on
-    return True
-
-
-def _is_moderation_int_errors(ex):
-    errs = ex.errors()
-    return all(map(_is_moderation_int_error, errs))
-
+    def parse(cls, chat_obj: str | dict) -> "ChatItem":
+        return _parse_chat_item(chat_obj)
 
-def _parse_chat(chat_text: str) -> ChatItem:
-    try:
-        return ChatItem.model_validate_json(chat_text)
-    except ValidationError as ex:
-        if _is_moderation_int_errors(ex):
-            logger.warning("Failed to parse ChatItem, fallback to old version with `Moderation:int`")
-        else:
-            logger.error(f"Failed to parse: {ex}")
 
-    # old version
-    chat_dict = json.loads(chat_text)
-    for rep in chat_dict["InnerContext"]["Replicas"]:
-        rep["Moderation"] = "OK"
+def _parse_chat_item(chat_obj: str | dict) -> ChatItem:
+    if isinstance(chat_obj, dict):
+        return ChatItem.model_validate(chat_obj)
 
-    return ChatItem.model_validate(chat_dict)
+    return ChatItem.model_validate_json(chat_obj)

{agi_med_common-5.0.15 → agi_med_common-5.0.17}/src/agi_med_common/models/widget.py

@@ -1,8 +1,10 @@
-from typing import List, Self
+from typing import List, Self, Literal
+
 from pydantic import BaseModel, model_validator
 
 
 class Widget(BaseModel):
+    type: Literal["widget"] = "widget"
     buttons: List[List[str]] | None = None
     ibuttons: List[List[str]] | None = None
 

agi_med_common-5.0.17/src/agi_med_common/type_union.py

@@ -0,0 +1,63 @@
+from typing import Any, Union, Annotated, get_args, get_origin, Literal
+
+from pydantic import Discriminator, BaseModel, Tag
+
+
+class _TypeUnionMeta(type):
+    def __getitem__(cls, types):
+        if not isinstance(types, tuple):
+            types = (types,)
+
+        for tp in types:
+            if not isinstance(tp, type) or not issubclass(tp, BaseModel):
+                raise ValueError(f"Type {tp} must derived from BaseModel")
+
+        # Create tagged union
+        tagged_types = []
+        type_map = {}
+        type_names = set()
+        for t in types:
+            type_annot = t.__annotations__.get("type")
+            if not type_annot or get_origin(type_annot) != Literal:
+                raise ValueError(f"Type {t} must have a 'type' Literal[..] field for discrimination")
+            type_name = get_args(type_annot)[0]
+            type_map[type_name] = t
+            tagged_types.append(Annotated[t, Tag(type_name)])
+            type_names.add(type_name)
+
+        # Create the union type
+        union_type = Union[tuple(tagged_types)]
+
+        # Create discriminator function
+        def type_discriminator(v: Any) -> str:
+            if isinstance(v, types):
+                return v.type
+            if isinstance(v, dict):
+                tp = v.get("type")
+                return tp if tp in type_names else None
+            return None
+
+        # Add discriminator
+        return Annotated[union_type, Discriminator(type_discriminator)]
+
+
+class TypeUnion(metaclass=_TypeUnionMeta):
+    """
+    Wrapper around Union which derive type from field 'type' for effective deserialization via TypeAdapter.
+
+    Usage example:
+    ```
+    class BaseFruit(BaseModel):
+        type: str
+
+    class Apple(BaseFruit):
+        type: Literal['apple'] = 'apple'
+
+    class Orange(BaseFruit):
+        type: Literal['orange'] = 'orange'
+
+    Fruit = TypeUnion[Apple, Orange]
+    ```
+    """
+
+    pass

{agi_med_common-5.0.15 → agi_med_common-5.0.17}/src/agi_med_common/utils.py

@@ -3,6 +3,7 @@ import json
 import os
 from datetime import datetime
 from pathlib import Path
+from typing import Iterable
 
 
 def make_session_id() -> str:
@@ -36,17 +37,20 @@ def try_parse_int(text: str) -> int | None:
     except (ValueError, TypeError):
         return None
 
+
 def try_parse_float(text: str) -> float | None:
     try:
         return float(text)
     except (ValueError, TypeError):
         return None
 
+
 def try_parse_bool(v: str | bool) -> bool:
     if isinstance(v, bool):
         return v
     return v.lower() in ("yes", "true", "t", "1")
 
+
 def pretty_line(text: str, cut_count: int = 100) -> str:
     if len(text) > 100:
         text_cut = text[:cut_count]
@@ -56,3 +60,10 @@ def pretty_line(text: str, cut_count: int = 100) -> str:
         text_pretty = text
     text_pretty = text_pretty.replace("\n", "\\n")
     return text_pretty
+
+
+def first_nonnull(obj: Iterable):
+    for elem in obj:
+        if elem:
+            return elem
+    return None

{agi_med_common-5.0.15 → agi_med_common-5.0.17}/src/agi_med_common.egg-info/PKG-INFO

@@ -1,12 +1,11 @@
 Metadata-Version: 2.4
 Name: agi_med_common
-Version: 5.0.15
+Version: 5.0.17
 Summary: Сommon for agi-med team
 Author: AGI-MED-TEAM
 Requires-Python: >=3.11
 Description-Content-Type: text/markdown
 Requires-Dist: pydantic<=3.0.0,>=2.9.2
-Requires-Dist: loguru~=0.7.2
 Requires-Dist: tqdm==4.67.*
 
 # agi-med-common

{agi_med_common-5.0.15 → agi_med_common-5.0.17}/src/agi_med_common.egg-info/SOURCES.txt

@@ -1,9 +1,9 @@
 README.md
 pyproject.toml
-requirements.txt
 src/agi_med_common/__init__.py
 src/agi_med_common/file_storage.py
 src/agi_med_common/parallel_map.py
+src/agi_med_common/type_union.py
 src/agi_med_common/utils.py
 src/agi_med_common/validators.py
 src/agi_med_common/xml_parser.py
@@ -13,15 +13,13 @@ src/agi_med_common.egg-info/dependency_links.txt
 src/agi_med_common.egg-info/requires.txt
 src/agi_med_common.egg-info/top_level.txt
 src/agi_med_common/api/chat_manager_api.py
+src/agi_med_common/api/classifier_api.py
 src/agi_med_common/api/content_interpreter_api.py
 src/agi_med_common/api/content_interpreter_remote_api.py
-src/agi_med_common/api/critic_api.py
-src/agi_med_common/api/text_processor_api.py
-src/agi_med_common/logger/__init__.py
-src/agi_med_common/logger/log_level_enum.py
-src/agi_med_common/logger/logger.py
+src/agi_med_common/api/text_generator_api.py
 src/agi_med_common/models/__init__.py
 src/agi_med_common/models/_base.py
+src/agi_med_common/models/chat.py
 src/agi_med_common/models/chat_item.py
 src/agi_med_common/models/enums.py
 src/agi_med_common/models/tracks.py

{agi_med_common-5.0.15 → agi_med_common-5.0.17}/src/agi_med_common.egg-info/requires.txt

@@ -1,3 +1,2 @@
 pydantic<=3.0.0,>=2.9.2
-loguru~=0.7.2
 tqdm==4.67.*

agi_med_common-5.0.15/requirements.txt

@@ -1,3 +0,0 @@
-pydantic>=2.9.2,<=3.0.0
-loguru~=0.7.2
-tqdm==4.67.*

agi_med_common-5.0.15/src/agi_med_common/api/critic_api.py

@@ -1,6 +0,0 @@
-from agi_med_common.models import ChatItem
-
-
-class CriticAPI:
-    def evaluate(self, text: str, chat: ChatItem | None = None, request_id: str = "") -> float:
-        raise NotImplementedError

agi_med_common-5.0.15/src/agi_med_common/api/text_processor_api.py

@@ -1,6 +0,0 @@
-from agi_med_common.models import ChatItem
-
-
-class TextProcessorAPI:
-    def process(self, text: str, chat: ChatItem | None = None, request_id: str = "") -> str:
-        raise NotImplementedError

agi_med_common-5.0.15/src/agi_med_common/logger/__init__.py

@@ -1,2 +0,0 @@
-from .log_level_enum import LogLevelEnum
-from .logger import logger_init, log_llm_error, logger

agi_med_common-5.0.15/src/agi_med_common/logger/log_level_enum.py

@@ -1,15 +0,0 @@
-from enum import StrEnum, auto
-
-
-class LogLevelEnum(StrEnum):
-    @staticmethod
-    def _generate_next_value_(name: str, start: int, count: int, last_values: list[str]) -> str:
-        return name.upper()
-
-    TRACE = auto()  # logger.trace()
-    DEBUG = auto()  # logger.debug()
-    INFO = auto()  # logger.info()
-    SUCCESS = auto()  # logger.success()
-    WARNING = auto()  # logger.warning()
-    ERROR = auto()  # logger.error()
-    CRITICAL = auto()  # logger.critical()

agi_med_common-5.0.15/src/agi_med_common/logger/logger.py

@@ -1,28 +0,0 @@
-import sys
-
-from loguru import logger
-
-from . import LogLevelEnum
-
-
-def logger_init(log_level: LogLevelEnum) -> None:
-    logger.remove()
-    extra = {"request_id": "SYSTEM_LOG"}
-    format_ = "{time:DD-MM-YYYY HH:mm:ss} | <level>{level: <8}</level> | {extra[request_id]}"
-    format_ = f"{format_} | <level>{{message}}</level>"
-    logger.add(sys.stdout, colorize=True, format=format_, level=log_level)
-    logger.configure(extra=extra)
-
-
-def log_llm_error(
-    text: str | None = None,
-    vector: list[float] | None = None,
-    model: str = "gigachat",
-) -> None:
-    if text is not None and not text:
-        logger.error(f"No response from {model}!!!")
-        return None
-    if vector is not None and all(not item for item in vector):
-        logger.error(f"No response from {model} encoder!!!")
-        return None
-    return None