solace-agent-mesh 0.0.1__py3-none-any.whl → 0.1.0__py3-none-any.whl

solace_agent_mesh/services/file_service/file_utils.py

@@ -0,0 +1,322 @@
+# Add utility functions for upload CSV that automatically creates the number of row and data types
+import csv
+from typing import Tuple, Union
+import json
+
+from .file_service_constants import FS_PROTOCOL, INDENT_SIZE
+
+
+class Types:
+    """
+    Type constants definition to be used when representing the schema.
+    """
+
+    INT = "int"
+    FLOAT = "float"
+    BOOL = "bool"
+    STR = "str"
+    ARRAY = "array"
+    OBJECT = "object"
+    NULL = "null"
+
+    @property
+    def PRIMITIVES():
+        return [Types.INT, Types.FLOAT, Types.BOOL, Types.STR]
+
+
+def get_str_type(value: str) -> str:
+    """
+    Get the type of a string value.
+
+    Parameters:
+    - value (str): The string value.
+
+    Returns:
+    - str: The type of the value.
+    """
+    if value.isdigit():
+        return Types.INT
+    elif value.replace(".", "", 1).isdigit():
+        return Types.FLOAT
+    else:
+        return Types.STR
+
+
+def convert_dict_to_yaml(schema: dict, indent_size=INDENT_SIZE, _level=0) -> str:
+    """
+    Convert a dictionary schema to a YAML string.
+
+    Parameters:
+    - schema (dict): The dictionary schema.
+
+    Returns:
+    - str: The YAML string.
+    """
+    yaml_str = ""
+    indent = indent_size * _level
+
+    for key, value in schema.items():
+        if isinstance(value, dict):
+            value_str = convert_dict_to_yaml(value, indent_size, _level + 1)
+            yaml_str += f"{space(indent)}{key}:\n{value_str}\n"
+        else:
+            yaml_str += f"{space(indent)}{key}: {value}\n"
+    return yaml_str.rstrip()
+
+
+def space(size=INDENT_SIZE):
+    """
+    Get a string of spaces.
+    """
+    if size == 0:
+        return ""
+    return " " * size
+
+
+def extract_csv_schema_and_shape(csv_file: Union[str, bytes]) -> Tuple[dict, Tuple[int, int]]:
+    """
+    Get the schema and shape of a CSV file.
+
+    Parameters:
+    - csv_file (str): The CSV file content.
+    - header_row (str): The header row of the CSV file.
+
+    Returns:
+    - str: The schema of the CSV file.
+    - st: The shape of the CSV file.
+    """
+    if type(csv_file) == bytes:
+        csv_file = csv_file.decode("utf-8")
+
+    # Parse the CSV string
+    csv_lines = csv_file.splitlines()
+    csv_reader = csv.DictReader(csv_lines)
+    columns = csv_reader.fieldnames
+    listed_csv = list(csv_reader)
+    shape = (len(listed_csv), len(columns))
+    schema = {}
+    for col in columns:
+        # deriving type from the top 10 rows
+        col_type = None
+        for i, row in enumerate(listed_csv):
+            if i >= 10:
+                break
+            curr_type = get_str_type(row[col])
+            if col_type is None:
+                col_type = curr_type
+            elif col_type != curr_type:
+                # multiple types, set as str
+                col_type = Types.STR
+                break
+        schema[col] = {"type": col_type}
+
+    return schema, shape
+
+
+def dict_to_schema(input_dict: dict) -> dict:
+    """
+    Convert a dictionary to a JSON schema.
+
+    Parameters:
+    - input_dict (dict): The input dictionary.
+
+    Returns:
+    - dict: The JSON schema.
+    """
+
+    def infer_type(value):
+        if isinstance(value, str):
+            return Types.STR
+        elif isinstance(value, int):
+            return Types.INT
+        elif isinstance(value, float):
+            return Types.FLOAT
+        elif isinstance(value, bool):
+            return Types.BOOL
+        elif isinstance(value, list):
+            if len(value) > 0:
+                return {"type": Types.ARRAY, "items": infer_type(value[0])}
+            else:
+                return {"type": Types.ARRAY, "items": {}}
+        elif isinstance(value, dict):
+            return {"type": Types.OBJECT, "properties": infer_properties(value)}
+        else:
+            return Types.NULL
+
+    def infer_properties(input_dict):
+        return {
+            key: (
+                {"type": infer_type(value)}
+                if isinstance(infer_type(value), str)
+                else infer_type(value)
+            )
+            for key, value in input_dict.items()
+        }
+
+    if type(input_dict) is list:
+        input_dict = input_dict[0]
+        json_schema = {
+            "type": Types.ARRAY,
+            "items": {"type": Types.OBJECT, "properties": infer_properties(input_dict)},
+        }
+    else:
+        json_schema = {"type": Types.OBJECT, "properties": infer_properties(input_dict)}
+
+    return json_schema
+
+
+def get_dict_array_shape(
+    data: Union[dict, list], max_level=2, _path="", _result=None, _level=0
+):
+    """
+    Get the count of each array in a dictionary schema.
+    """
+    if _result is None:
+        _result = []
+
+    # If it's a list, record its length and traverse each element
+    if isinstance(data, list):
+        if _path:
+            _result.append([_path, len(data)])
+        else:
+            _result.append(["top level", len(data)])
+        if _level < max_level:
+            get_dict_array_shape(data[0], max_level, f"{_path}[*]", _result, _level + 1)
+
+    # If it's a dictionary, traverse each key-value pair
+    elif isinstance(data, dict):
+        for key, value in data.items():
+            new_path = f"{_path}.{key}" if _path else key
+            if _level < max_level:
+                get_dict_array_shape(value, max_level, new_path, _result, _level + 1)
+
+    return _result
+
+
+def get_csv_schema_and_shape(schema: dict, shape: Tuple[int, int]) -> Tuple[str, str]:
+    """
+    Get the formatted schema and shape of a CSV file.
+
+    Parameters:
+    - schema (dict): The schema of the CSV file.
+    - shape (Tuple[int, int]): The shape of the CSV file.
+
+    Returns:
+    - str: The formatted schema of the CSV file.
+    - Tuple[int, int]: The shape of the CSV file.
+    """
+    shape_str = f"{shape[0]} rows x {shape[1]} columns"
+    schema_dict = {
+        "type": Types.ARRAY,
+        "items": {"type": Types.OBJECT, "properties": schema},
+    }
+    schema_str = convert_dict_to_yaml(schema_dict)
+    return schema_str, shape_str
+
+
+def get_json_schema_and_shape(file: Union[bytes, str]) -> Tuple[str, str]:
+    try:
+        if type(file) == bytes:
+            file = file.decode("utf-8")
+        json_data = json.loads(file)
+        schema = dict_to_schema(json_data)
+        schema_str = convert_dict_to_yaml(schema)
+        shape_str = None
+
+        def get_row_text(count):
+            return f"{count} elements" if count > 1 else f"{count} element"
+
+        array_lengths = get_dict_array_shape(json_data)
+        if array_lengths:
+            shape_str = "\n".join(
+                [f"{key}: {get_row_text(val)}" for key, val in array_lengths]
+            )
+        return schema_str, shape_str
+    except Exception as e:
+        return None, None
+
+
+def get_file_schema_and_shape(file: bytes, metadata: dict) -> Tuple[str, str]:
+    """
+    Get the schema and shape of a file.
+
+    Parameters:
+    - file (bytes): The file content as bytes.
+    - metadata (dict): The file metadata.
+
+    Returns:
+    - str: The schema of the file. None if cannot be determined.
+    - str: The shape of the file. None if cannot be determined.
+    """
+    if metadata.get("mime_type") == "text/csv":
+        schema, shape = extract_csv_schema_and_shape(file)
+        schema_str, shape_str = get_csv_schema_and_shape(schema, shape)
+        return schema_str, shape_str
+    elif metadata.get("mime_type") == "application/json":
+        schema, shape = get_json_schema_and_shape(file)
+        return schema, shape
+    else:
+        return None, None
+
+
+def recursive_file_resolver(
+    obj: Union[dict, str, list],
+    resolver: callable,
+    session_id: str,
+    force_resolve: bool = False,
+):
+    """
+    Recursively resolve all the file URLs in the input data.
+
+    Parameters:
+    - obj (dict|str|list): The input data.
+    - resolver (callable): The resolver function.
+    - session_id (str): The session ID.
+    - force_resolve (bool): Resolve all the URLs regardless of the 'resolve' query parameter.
+
+    Returns:
+    - dict|str|list: The object with resolved URLs.
+    """
+    if type(obj) == str:
+        return resolver(obj, session_id, force_resolve)
+
+    keys = []
+    if type(obj) == list:
+        keys = range(len(obj))
+    elif type(obj) == dict:
+        keys = obj.keys()
+
+    for key in keys:
+        if isinstance(obj[key], dict):
+            obj[key] = recursive_file_resolver(
+                obj[key], resolver, session_id, force_resolve
+            )
+        elif isinstance(obj[key], list):
+            obj[key] = [
+                (
+                    recursive_file_resolver(item, resolver, session_id, force_resolve)
+                    if isinstance(item, dict)
+                    else item
+                )
+                for item in obj[key]
+            ]
+        elif isinstance(obj[key], str):
+            obj[key] = resolver(obj[key], session_id, force_resolve)
+    return obj
+
+
+def starts_with_fs_url(url: str) -> bool:
+    """
+    Check if the URL starts with the file service protocol.
+
+    Parameters:
+    - url (str): The URL.
+
+    Returns:
+    - bool: True if the URL starts with the file service protocol.
+    """
+    return (
+        url.startswith(f"{FS_PROTOCOL}://")
+        or url.startswith(f"<url>{FS_PROTOCOL}://")
+        or url.startswith(f"<url> {FS_PROTOCOL}://")
+    )

solace_agent_mesh/services/file_service/transformers/__init__.py

@@ -0,0 +1,5 @@
+TRANSFORMERS = []
+
+__all__ = [
+    "TRANSFORMERS"
+]

solace_agent_mesh/services/history_service/__init__.py

@@ -0,0 +1,3 @@
+from .history_service import HistoryService
+
+__all__ = ["HistoryService"]

solace_agent_mesh/services/history_service/history_providers/base_history_provider.py

@@ -0,0 +1,78 @@
+from abc import ABC, abstractmethod
+from typing import Union
+
+class BaseHistoryProvider(ABC):
+
+    def __init__(self, config=None):
+        self.config = config
+        self.max_turns = self.config.get("max_turns")
+        self.max_characters = self.config.get("max_characters")
+        self.enforce_alternate_message_roles = self.config.get("enforce_alternate_message_roles")
+
+    @abstractmethod
+    def store_history(self, session_id: str, role: str, content: Union[str, dict]):
+        """
+        Store a new entry in the history.
+
+        :param session_id: The session identifier.
+        :param role: The role of the entry to be stored in the history.
+        :param content: The content of the entry to be stored in the history.
+        """
+        raise NotImplementedError("Method not implemented")
+
+    @abstractmethod
+    def get_history(self, session_id: str):
+        """
+        Retrieve the entire history.
+
+        :param session_id: The session identifier.
+        :return: The complete history.
+        """
+        raise NotImplementedError("Method not implemented")
+
+    @abstractmethod
+    def store_file(self, session_id: str, file: dict):
+        """
+        Store a file in the history.
+
+        :param session_id: The session identifier.
+        :param file: The file metadata to be stored in the history.
+        """
+        raise NotImplementedError("Method not implemented")
+
+    @abstractmethod
+    def get_files(self, session_id: str):
+        """
+        Retrieve the files for a session.
+
+        :param session_id: The session identifier.
+        :return: The files for the session.
+        """
+        raise NotImplementedError("Method not implemented")
+    
+    @abstractmethod
+    def get_session_meta(self, session_id: str):
+        """
+        Retrieve the session metadata.
+
+        :param session_id: The session identifier.
+        :return: The session metadata.
+        """
+        raise NotImplementedError("Method not implemented")
+
+    @abstractmethod
+    def clear_history(self, session_id: str, keep_levels=0):
+        """
+        Clear the history and files, optionally keeping a specified number of recent entries.
+
+        :param session_id: The session identifier.
+        :param keep_levels: Number of most recent history entries to keep. Default is 0 (clear all).
+        """
+        raise NotImplementedError("Method not implemented")
+
+    @abstractmethod
+    def get_all_sessions(self) -> list[str]:
+        """
+        Retrieve all session identifiers.
+        """
+        raise NotImplementedError("Method not implemented")

solace_agent_mesh/services/history_service/history_providers/memory_history_provider.py

@@ -0,0 +1,167 @@
+import time
+from .base_history_provider import BaseHistoryProvider
+
+
+class MemoryHistoryProvider(BaseHistoryProvider):
+    def __init__(self, config=None):
+        super().__init__(config)
+        self.history = {}
+
+    def store_history(self, session_id: str, role: str, content: str | dict):
+        """
+        Store a new entry in the history.
+
+        :param session_id: The session identifier.
+        :param history_entry: The entry to be stored in the history.
+        """
+
+        if session_id not in self.history:
+            self.history[session_id] = {
+                "history": [],
+                "files": [],
+                "last_active_time": time.time(),
+                "num_characters": 0,
+                "num_turns": 0,
+            }
+
+        # Check if adding another entry would exceed the max_turns
+        if self.history[session_id]["num_turns"] == self.max_turns:
+            # Remove the oldest entry
+            oldest_entry = self.history[session_id]["history"].pop(0)
+            # Subtract the length of the oldest entry from the total length
+            self.history[session_id]["num_characters"] -= len(
+                str(oldest_entry["content"])
+            )
+            self.history[session_id]["num_turns"] -= 1
+
+        if (
+            self.enforce_alternate_message_roles
+            and self.history[session_id]["num_turns"] > 0
+            # Check if the last entry was by the same role
+            and self.history[session_id]["history"]
+            and self.history[session_id]["history"][-1]["role"] == role
+        ):
+            # Append to last entry
+            self.history[session_id]["history"][-1]["content"] += content
+        else:
+            # Add the new entry
+            self.history[session_id]["history"].append(
+                {"role": role, "content": content}
+            )
+            # Update the number of turns
+            self.history[session_id]["num_turns"] += 1
+
+        # Update the length
+        self.history[session_id]["num_characters"] += len(str(content))
+        # Update the last active time
+        self.history[session_id]["last_active_time"] = time.time()
+
+        # Check if we have exceeded max_characters
+        if self.max_characters:
+            while (
+                self.history[session_id]["num_characters"] > self.max_characters
+                and self.history[session_id]["num_turns"] > 0
+            ):
+                # Remove the oldest entry
+                oldest_entry = self.history[session_id]["history"].pop(0)
+                # Subtract the length of the oldest entry from the total length
+                self.history[session_id]["num_characters"] -= len(
+                    str(oldest_entry["content"])
+                )
+                self.history[session_id]["num_turns"] -= 1
+
+    def get_history(self, session_id: str):
+        """
+        Retrieve the entire history for a session.
+
+        :param session_id: The session identifier.
+        :return: The complete history for the session.
+        """
+        if session_id not in self.history:
+            return []
+        return self.history.get(session_id)["history"]
+
+    def store_file(self, session_id: str, file: dict):
+        """
+        Store a file in the history.
+
+        :param session_id: The session identifier.
+        :param file: The file metadata to be stored in the history.
+        """
+        if session_id not in self.history:
+            self.history[session_id] = {
+                "history": [],
+                "files": [],
+                "last_active_time": time.time(),
+                "num_characters": 0,
+                "num_turns": 0,
+            }
+        self.history[session_id]["last_active_time"] = time.time()
+
+        # Check duplicate
+        for f in self.history[session_id]["files"]:
+            if f.get("url") and f.get("url") == file.get("url"):
+                return
+
+        self.history[session_id]["files"].append(file)
+
+    def get_files(self, session_id: str):
+        """
+        Retrieve the files for a session.
+
+        :param session_id: The session identifier.
+        :return: The files for the session.
+        """
+        if session_id not in self.history:
+            return []
+        files = []
+        current_time = time.time()
+        all_files = self.history.get(session_id)["files"].copy()
+        for file in all_files:
+            expiration_timestamp = file.get("expiration_timestamp")
+            if expiration_timestamp and current_time > expiration_timestamp:
+                self.history[session_id]["files"].remove(file)
+                continue
+            files.append(file)
+        return files
+
+    def clear_history(self, session_id: str, keep_levels=0):
+        """
+        Clear the history for a session, optionally keeping a specified number of recent entries.
+
+        :param session_id: The session identifier.
+        :param keep_levels: Number of most recent history entries to keep. Default is 0 (clear all).
+        """
+        if session_id in self.history:
+            if keep_levels <= 0:
+                del self.history[session_id]
+            else:
+                self.history[session_id]["history"] = self.history[session_id][
+                    "history"
+                ][-keep_levels:]
+                # Recalculate the length and num_turns
+                self.history[session_id]["num_characters"] = sum(
+                    len(str(entry)) for entry in self.history[session_id]["history"]
+                )
+                self.history[session_id]["num_turns"] = len(
+                    self.history[session_id]["history"]
+                )
+
+    def get_session_meta(self, session_id: str):
+        """
+        Retrieve the session metadata.
+
+        :param session_id: The session identifier.
+        :return: The session metadata.
+        """
+        if session_id in self.history:
+            session = self.history[session_id]
+            return {
+                "num_characters": session["num_characters"],
+                "num_turns": session["num_turns"],
+                "last_active_time": session["last_active_time"],
+            }
+        return None
+
+    def get_all_sessions(self) -> list[str]:
+        return list(self.history.keys())