doc changes

- Add generic CredentialStore abstraction for credential persistence
- Introduce pickle- and Redis-backed credential store implementations
- Refactor Google OAuth provider to delegate persistence to CredentialStore
- Make auth providers generic over credential type for stricter contracts
- Update package documentation to reflect credential lifecycle vs persistence split
- Add credentials module to public API surface
- Harden .gitignore to exclude credential and token artifacts

This release removes node-local persistence assumptions, enables
distributed-safe authentication, and formalizes credential storage as
a first-class extension point.

.drone.yml

@@ -0,0 +1,129 @@
+---
+kind: pipeline
+type: docker
+name: build-and-publish-pypi
+
+platform:
+  os: linux
+  arch: arm64
+
+workspace:
+  path: /drone/src
+
+steps:
+  - name: check-version
+    image: curlimages/curl:latest
+    environment:
+      PIP_REPO_URL:
+        from_secret: PIP_REPO_URL
+      PIP_USERNAME:
+        from_secret: PIP_USERNAME
+      PIP_PASSWORD:
+        from_secret: PIP_PASSWORD
+    commands:
+      - PACKAGE_NAME=$(grep -E '^name\s*=' pyproject.toml | head -1 | cut -d'"' -f2)
+      - VERSION=$(grep -E '^version\s*=' pyproject.toml | head -1 | cut -d'"' -f2)
+      - echo "🔍 Checking if $PACKAGE_NAME==$VERSION exists on $PIP_REPO_URL ..."
+      - |
+        if curl -fsSL -u "$PIP_USERNAME:$PIP_PASSWORD" "$PIP_REPO_URL/simple/$PACKAGE_NAME/" | grep -q "$VERSION"; then
+          echo "✅ $PACKAGE_NAME==$VERSION already exists — skipping build."
+          exit 78
+        else
+          echo "🆕 New version detected: $PACKAGE_NAME==$VERSION"
+        fi
+
+  - name: build-package
+    image: python:3.13-slim
+    commands:
+      - pip install --upgrade pip build
+      - echo "📦 Building Python package..."
+      - python -m build
+      - ls -l dist
+
+  - name: upload-to-private-pypi
+    image: python:3.13-slim
+    environment:
+      PIP_REPO_URL:
+        from_secret: PIP_REPO_URL
+      PIP_USERNAME:
+        from_secret: PIP_USERNAME
+      PIP_PASSWORD:
+        from_secret: PIP_PASSWORD
+    commands:
+      - pip install --upgrade twine
+      - echo "🚀 Uploading to private PyPI at $PIP_REPO_URL ..."
+      - |
+        twine upload \
+          --repository-url "$PIP_REPO_URL" \
+          -u "$PIP_USERNAME" \
+          -p "$PIP_PASSWORD" \
+          dist/*
+
+trigger:
+  event:
+    - tag
+
+---
+kind: pipeline
+type: docker
+name: backfill-pypi-from-tags
+
+platform:
+  os: linux
+  arch: arm64
+
+workspace:
+  path: /drone/src
+
+steps:
+  - name: fetch-tags
+    image: alpine/git
+    commands:
+      - git fetch --tags --force
+
+  - name: build-and-upload-missing
+    image: python:3.13-slim
+    environment:
+      PIP_REPO_URL:
+        from_secret: PIP_REPO_URL
+      PIP_USERNAME:
+        from_secret: PIP_USERNAME
+      PIP_PASSWORD:
+        from_secret: PIP_PASSWORD
+    commands:
+      - apt-get update
+      - apt-get install -y git curl ca-certificates
+      - pip install --upgrade pip build twine
+      - |
+        set -e
+
+        PACKAGE_NAME=$(grep -E '^name\s*=' pyproject.toml | cut -d'"' -f2)
+        echo "📦 Package: $PACKAGE_NAME"
+
+        for TAG in $(git tag --sort=version:refname); do
+          VERSION="$TAG"
+          echo "🔁 Version: $VERSION"
+
+          if curl -fsSL -u "$PIP_USERNAME:$PIP_PASSWORD" \
+            "$PIP_REPO_URL/simple/$PACKAGE_NAME/" | grep -q "$VERSION"; then
+            echo "⏭️  Exists, skipping"
+            continue
+          fi
+
+          git checkout --force "$TAG"
+
+          echo "🏗️  Building $VERSION"
+          rm -rf dist
+          python -m build
+
+          echo "⬆️  Uploading $VERSION"
+          twine upload \
+            --repository-url "$PIP_REPO_URL" \
+            -u "$PIP_USERNAME" \
+            -p "$PIP_PASSWORD" \
+            dist/*
+        done
+
+trigger:
+  event:
+    - custom

.gitignore

@@ -38,3 +38,8 @@ Thumbs.db
 *.swo
 *~
 *.tmp
+
+# Credentials
+client_secret_*.json
+token.pickle
+credentials*.json

docs/mail_intake/adapters/base.md

@@ -0,0 +1,3 @@
+# Base
+
+::: mail_intake.adapters.base

docs/mail_intake/adapters/gmail.md

@@ -0,0 +1,3 @@
+# Gmail
+
+::: mail_intake.adapters.gmail

docs/mail_intake/adapters/index.md

@@ -0,0 +1,3 @@
+# Adapters
+
+::: mail_intake.adapters

docs/mail_intake/auth/base.md

@@ -0,0 +1,3 @@
+# Base
+
+::: mail_intake.auth.base

docs/mail_intake/auth/google.md

@@ -0,0 +1,3 @@
+# Google
+
+::: mail_intake.auth.google

docs/mail_intake/auth/index.md

@@ -0,0 +1,3 @@
+# Auth
+
+::: mail_intake.auth

docs/mail_intake/config.md

@@ -0,0 +1,3 @@
+# Config
+
+::: mail_intake.config

docs/mail_intake/exceptions.md

@@ -0,0 +1,3 @@
+# Exceptions
+
+::: mail_intake.exceptions

docs/mail_intake/index.md

@@ -0,0 +1,3 @@
+# Mail Intake
+
+::: mail_intake

docs/mail_intake/ingestion/index.md

@@ -0,0 +1,3 @@
+# Ingestion
+
+::: mail_intake.ingestion

docs/mail_intake/ingestion/reader.md

@@ -0,0 +1,3 @@
+# Reader
+
+::: mail_intake.ingestion.reader

docs/mail_intake/models/index.md

@@ -0,0 +1,3 @@
+# Models
+
+::: mail_intake.models

docs/mail_intake/models/message.md

@@ -0,0 +1,3 @@
+# Message
+
+::: mail_intake.models.message

docs/mail_intake/models/thread.md

@@ -0,0 +1,3 @@
+# Thread
+
+::: mail_intake.models.thread

docs/mail_intake/parsers/body.md

@@ -0,0 +1,3 @@
+# Body
+
+::: mail_intake.parsers.body

docs/mail_intake/parsers/headers.md

@@ -0,0 +1,3 @@
+# Headers
+
+::: mail_intake.parsers.headers

docs/mail_intake/parsers/index.md

@@ -0,0 +1,3 @@
+# Parsers
+
+::: mail_intake.parsers

docs/mail_intake/parsers/subject.md

@@ -0,0 +1,3 @@
+# Subject
+
+::: mail_intake.parsers.subject

mail_intake/__init__.py

@@ -0,0 +1,136 @@
+"""
+Mail Intake — provider-agnostic, read-only email ingestion framework.
+
+Mail Intake is a **contract-first library** designed to ingest, parse, and
+normalize email data from external providers (such as Gmail) into clean,
+provider-agnostic domain models.
+
+The library is intentionally structured around clear layers, each exposed
+as a first-class module at the package root:
+
+- adapters: provider-specific access (e.g. Gmail)
+- auth: authentication providers and credential lifecycle management
+- credentials: credential persistence abstractions and implementations
+- parsers: extraction and normalization of message content
+- ingestion: orchestration and high-level ingestion workflows
+- models: canonical, provider-agnostic data representations
+- config: explicit global configuration
+- exceptions: library-defined error hierarchy
+
+The package root acts as a **namespace**, not a facade. Consumers are
+expected to import functionality explicitly from the appropriate module.
+
+----------------------------------------------------------------------
+Installation
+----------------------------------------------------------------------
+
+Install using pip:
+
+    pip install mail-intake
+
+Or with Poetry:
+
+    poetry add mail-intake
+
+Mail Intake is pure Python and has no runtime dependencies beyond those
+required by the selected provider (for example, Google APIs for Gmail).
+
+----------------------------------------------------------------------
+Basic Usage
+----------------------------------------------------------------------
+
+Minimal Gmail ingestion example (local development):
+
+    from mail_intake.ingestion import MailIntakeReader
+    from mail_intake.adapters import MailIntakeGmailAdapter
+    from mail_intake.auth import MailIntakeGoogleAuth
+    from mail_intake.credentials import PickleCredentialStore
+
+    store = PickleCredentialStore(path="token.pickle")
+
+    auth = MailIntakeGoogleAuth(
+        credentials_path="credentials.json",
+        store=store,
+        scopes=["https://www.googleapis.com/auth/gmail.readonly"],
+    )
+
+    adapter = MailIntakeGmailAdapter(auth_provider=auth)
+    reader = MailIntakeReader(adapter)
+
+    for message in reader.iter_messages("from:recruiter@example.com"):
+        print(message.subject, message.from_email)
+
+Iterating over threads:
+
+    for thread in reader.iter_threads("subject:Interview"):
+        print(thread.normalized_subject, len(thread.messages))
+
+----------------------------------------------------------------------
+Extensibility Model
+----------------------------------------------------------------------
+
+Mail Intake is designed to be extensible via **public contracts** exposed
+through its modules:
+
+- Users MAY implement their own mail adapters by subclassing
+  ``adapters.MailIntakeAdapter``
+- Users MAY implement their own authentication providers by subclassing
+  ``auth.MailIntakeAuthProvider[T]``
+- Users MAY implement their own credential persistence layers by
+  implementing ``credentials.CredentialStore[T]``
+
+Users SHOULD NOT subclass built-in adapter implementations. Built-in
+adapters (such as Gmail) are reference implementations and may change
+internally without notice.
+
+----------------------------------------------------------------------
+Public API Surface
+----------------------------------------------------------------------
+
+The supported public API consists of the following top-level modules:
+
+- mail_intake.ingestion
+- mail_intake.adapters
+- mail_intake.auth
+- mail_intake.credentials
+- mail_intake.parsers
+- mail_intake.models
+- mail_intake.config
+- mail_intake.exceptions
+
+Classes and functions should be imported explicitly from these modules.
+No individual symbols are re-exported at the package root.
+
+----------------------------------------------------------------------
+Design Guarantees
+----------------------------------------------------------------------
+
+- Read-only access: no mutation of provider state
+- Provider-agnostic domain models
+- Explicit configuration and dependency injection
+- No implicit global state or environment reads
+- Deterministic, testable behavior
+- Distributed-safe authentication design
+
+Mail Intake favors correctness, clarity, and explicitness over convenience
+shortcuts.
+"""
+
+
+from . import ingestion
+from . import adapters
+from . import auth
+from . import credentials
+from . import models
+from . import config
+from . import exceptions
+
+__all__ = [
+    "ingestion",
+    "adapters",
+    "auth",
+    "credentials",
+    "models",
+    "config",
+    "exceptions",
+]

mail_intake/adapters/__init__.py

@@ -0,0 +1,28 @@
+"""
+Mail provider adapter implementations for Mail Intake.
+
+This package contains **adapter-layer implementations** responsible for
+interfacing with external mail providers and exposing a normalized,
+provider-agnostic contract to the rest of the system.
+
+Adapters in this package:
+- Implement the `MailIntakeAdapter` interface
+- Encapsulate all provider-specific APIs and semantics
+- Perform read-only access to mail data
+- Return provider-native payloads without interpretation
+
+Provider-specific logic **must not leak** outside of adapter implementations.
+All parsings, normalizations, and transformations must be handled by downstream
+components.
+
+Public adapters exported from this package are considered the supported
+integration surface for mail providers.
+"""
+
+from .base import MailIntakeAdapter
+from .gmail import MailIntakeGmailAdapter
+
+__all__ = [
+    "MailIntakeAdapter",
+    "MailIntakeGmailAdapter",
+]

mail_intake/adapters/base.py

@@ -1,3 +1,14 @@
+"""
+Mail provider adapter contracts for Mail Intake.
+
+This module defines the **provider-agnostic adapter interface** used for
+read-only mail ingestion.
+
+Adapters encapsulate all provider-specific access logic and expose a
+minimal, normalized contract to the rest of the system. No provider-specific
+types or semantics should leak beyond implementations of this interface.
+"""
+
 from abc import ABC, abstractmethod
 from typing import Iterator, Dict, Any
 
@@ -6,19 +17,28 @@ class MailIntakeAdapter(ABC):
     """
     Base adapter interface for mail providers.
 
-    This interface defines the minimal contract required for
+    This interface defines the minimal contract required to:
-    read-only mail ingestion. No provider-specific concepts
+    - Discover messages matching a query
-    should leak beyond implementations of this class.
+    - Retrieve full message payloads
+    - Retrieve full thread payloads
+
+    Adapters are intentionally read-only and must not mutate provider state.
     """
 
     @abstractmethod
     def iter_message_refs(self, query: str) -> Iterator[Dict[str, str]]:
         """
-        Iterate over lightweight message references.
+        Iterate over lightweight message references matching a query.
 
-        Must yield dictionaries containing at least:
+        Implementations must yield dictionaries containing at least:
-        - message_id
+        - ``message_id``: Provider-specific message identifier
-        - thread_id
+        - ``thread_id``: Provider-specific thread identifier
+
+        Args:
+            query: Provider-specific query string used to filter messages.
+
+        Yields:
+            Dictionaries containing message and thread identifiers.
 
         Example yield:
             {
@@ -31,18 +51,26 @@ class MailIntakeAdapter(ABC):
     @abstractmethod
     def fetch_message(self, message_id: str) -> Dict[str, Any]:
         """
-        Fetch a full raw message by message_id.
+        Fetch a full raw message by message identifier.
 
-        Returns the provider-native message payload
+        Args:
-        (e.g., Gmail message JSON).
+            message_id: Provider-specific message identifier.
+
+        Returns:
+            Provider-native message payload
+            (e.g., Gmail message JSON structure).
         """
         raise NotImplementedError
 
     @abstractmethod
     def fetch_thread(self, thread_id: str) -> Dict[str, Any]:
         """
-        Fetch a full raw thread by thread_id.
+        Fetch a full raw thread by thread identifier.
 
-        Returns the provider-native thread payload.
+        Args:
+            thread_id: Provider-specific thread identifier.
+
+        Returns:
+            Provider-native thread payload.
         """
         raise NotImplementedError

mail_intake/adapters/gmail.py

@@ -1,3 +1,17 @@
+"""
+Gmail adapter implementation for Mail Intake.
+
+This module provides a **Gmail-specific implementation** of the
+`MailIntakeAdapter` contract.
+
+It is the only place in the codebase where:
+- `googleapiclient` is imported
+- Gmail REST API semantics are known
+- Low-level `.execute()` calls are made
+
+All Gmail-specific behavior must be strictly contained within this module.
+"""
+
 from typing import Iterator, Dict, Any
 
 from googleapiclient.discovery import build
@@ -12,12 +26,19 @@ class MailIntakeGmailAdapter(MailIntakeAdapter):
     """
     Gmail read-only adapter.
 
+    This adapter implements the `MailIntakeAdapter` interface using the
+    Gmail REST API. It translates the generic mail intake contract into
+    Gmail-specific API calls.
+
     This class is the ONLY place where:
     - googleapiclient is imported
     - Gmail REST semantics are known
     - .execute() is called
 
-    It must remain thin and dumb by design.
+    Design constraints:
+    - Must remain thin and imperative
+    - Must not perform parsing or interpretation
+    - Must not expose Gmail-specific types beyond this class
     """
 
     def __init__(
@@ -25,12 +46,29 @@ class MailIntakeGmailAdapter(MailIntakeAdapter):
         auth_provider: MailIntakeAuthProvider,
         user_id: str = "me",
     ):
+        """
+        Initialize the Gmail adapter.
+
+        Args:
+            auth_provider: Authentication provider capable of supplying
+                valid Gmail API credentials.
+            user_id: Gmail user identifier. Defaults to `"me"`.
+        """
         self._auth_provider = auth_provider
         self._user_id = user_id
         self._service = None
 
     @property
     def service(self):
+        """
+        Lazily initialize and return the Gmail API service client.
+
+        Returns:
+            Initialized Gmail API service instance.
+
+        Raises:
+            MailIntakeAdapterError: If the Gmail service cannot be initialized.
+        """
         if self._service is None:
             try:
                 creds = self._auth_provider.get_credentials()
@@ -45,11 +83,16 @@ class MailIntakeGmailAdapter(MailIntakeAdapter):
         """
         Iterate over message references matching the query.
 
+        Args:
+            query: Gmail search query string.
+
         Yields:
-        {
+            Dictionaries containing:
-            "message_id": "...",
+            - ``message_id``: Gmail message ID
-            "thread_id": "..."
+            - ``thread_id``: Gmail thread ID
-        }
+
+        Raises:
+            MailIntakeAdapterError: If the Gmail API returns an error.
         """
         try:
             request = (
@@ -79,6 +122,18 @@ class MailIntakeGmailAdapter(MailIntakeAdapter):
             ) from exc
 
     def fetch_message(self, message_id: str) -> Dict[str, Any]:
+        """
+        Fetch a full Gmail message by message ID.
+
+        Args:
+            message_id: Gmail message identifier.
+
+        Returns:
+            Provider-native Gmail message payload.
+
+        Raises:
+            MailIntakeAdapterError: If the Gmail API returns an error.
+        """
         try:
             return (
                 self.service.users()
@@ -92,6 +147,18 @@ class MailIntakeGmailAdapter(MailIntakeAdapter):
             ) from exc
 
     def fetch_thread(self, thread_id: str) -> Dict[str, Any]:
+        """
+        Fetch a full Gmail thread by thread ID.
+
+        Args:
+            thread_id: Gmail thread identifier.
+
+        Returns:
+            Provider-native Gmail thread payload.
+
+        Raises:
+            MailIntakeAdapterError: If the Gmail API returns an error.
+        """
         try:
             return (
                 self.service.users()

mail_intake/auth/__init__.py

@@ -0,0 +1,26 @@
+"""
+Authentication provider implementations for Mail Intake.
+
+This package defines the **authentication layer** used by mail adapters
+to obtain provider-specific credentials.
+
+It exposes:
+- A stable, provider-agnostic authentication contract
+- Concrete authentication providers for supported platforms
+
+Authentication providers:
+- Are responsible for credential acquisition and lifecycle management
+- Are intentionally decoupled from adapter logic
+- May be extended by users to support additional providers
+
+Consumers should depend on the abstract interface and use concrete
+implementations only where explicitly required.
+"""
+
+from .base import MailIntakeAuthProvider
+from .google import MailIntakeGoogleAuth
+
+__all__ = [
+    "MailIntakeAuthProvider",
+    "MailIntakeGoogleAuth",
+]

mail_intake/auth/base.py

@@ -1,20 +1,59 @@
-from abc import ABC, abstractmethod
-
-
-class MailIntakeAuthProvider(ABC):
 """
-    Abstract authentication provider.
+Authentication provider contracts for Mail Intake.
 
-    Mail adapters depend on this interface, not on concrete
+This module defines the **authentication abstraction layer** used by mail
-    OAuth or credential implementations.
+adapters to obtain provider-specific credentials.
+
+Authentication concerns are intentionally decoupled from adapter logic.
+Adapters depend only on this interface and must not be aware of how
+credentials are acquired, refreshed, or persisted.
+"""
+
+from abc import ABC, abstractmethod
+from typing import Generic, TypeVar
+
+T = TypeVar("T")
+
+
+class MailIntakeAuthProvider(ABC, Generic[T]):
+    """
+    Abstract base class for authentication providers.
+
+    This interface enforces a strict contract between authentication
+    providers and mail adapters by requiring providers to explicitly
+    declare the type of credentials they return.
+
+    Authentication providers encapsulate all logic required to:
+    - Acquire credentials from an external provider
+    - Refresh or revalidate credentials as needed
+    - Handle authentication-specific failure modes
+    - Coordinate with credential persistence layers where applicable
+
+    Mail adapters must treat returned credentials as opaque and
+    provider-specific, relying only on the declared credential type
+    expected by the adapter.
     """
 
     @abstractmethod
-    def get_credentials(self):
+    def get_credentials(self) -> T:
         """
-        Return provider-specific credentials object.
+        Retrieve valid, provider-specific credentials.
 
-        This method is synchronous by design and must either
+        This method is synchronous by design and represents the sole
-        return valid credentials or raise MailIntakeAuthError.
+        entry point through which adapters obtain authentication
+        material.
+
+        Implementations must either return credentials of the declared
+        type ``T`` that are valid at the time of return or raise an
+        authentication-specific exception.
+
+        Returns:
+            Credentials of type ``T`` suitable for immediate use by the
+            corresponding mail adapter.
+
+        Raises:
+            Exception:
+                An authentication-specific exception indicating that
+                credentials could not be obtained or validated.
         """
         raise NotImplementedError

mail_intake/auth/google.py

@@ -1,12 +1,28 @@
+"""
+Google authentication provider implementation for Mail Intake.
+
+This module provides a **Google OAuth–based authentication provider**
+used primarily for Gmail access.
+
+It encapsulates all Google-specific authentication concerns, including:
+- Credential loading and persistence
+- Token refresh handling
+- Interactive OAuth flow initiation
+- Coordination with a credential persistence layer
+
+No Google authentication details should leak outside this module.
+"""
+
 import os
-import pickle
 from typing import Sequence
 
 import google.auth.exceptions
 from google.auth.transport.requests import Request
 from google_auth_oauthlib.flow import InstalledAppFlow
+from google.oauth2.credentials import Credentials
 
 from mail_intake.auth.base import MailIntakeAuthProvider
+from mail_intake.credentials.store import CredentialStore
 from mail_intake.exceptions import MailIntakeAuthError
 
 
@@ -14,34 +30,62 @@ class MailIntakeGoogleAuth(MailIntakeAuthProvider):
     """
     Google OAuth provider for Gmail access.
 
-    Responsibilities:
+    This provider implements the `MailIntakeAuthProvider` interface using
-    - Load cached credentials from disk
+    Google's OAuth 2.0 flow and credential management libraries.
-    - Refresh expired tokens when possible
-    - Trigger interactive login only when strictly required
 
-    This class is synchronous and intentionally state-light.
+    Responsibilities:
+    - Load cached credentials from a credential store when available
+    - Refresh expired credentials when possible
+    - Initiate an interactive OAuth flow only when required
+    - Persist refreshed or newly obtained credentials via the store
+
+    This class is synchronous by design and maintains a minimal internal state.
     """
 
     def __init__(
         self,
         credentials_path: str,
-        token_path: str,
+        store: CredentialStore[Credentials],
         scopes: Sequence[str],
     ):
+        """
+        Initialize the Google authentication provider.
+
+        Args:
+            credentials_path:
+                Path to the Google OAuth client secrets file used to
+                initiate the OAuth 2.0 flow.
+
+            store:
+                Credential store responsible for persisting and
+                retrieving Google OAuth credentials.
+
+            scopes:
+                OAuth scopes required for Gmail access.
+        """
         self.credentials_path = credentials_path
-        self.token_path = token_path
+        self.store = store
         self.scopes = list(scopes)
 
-    def get_credentials(self):
+    def get_credentials(self) -> Credentials:
-        creds = None
+        """
+        Retrieve valid Google OAuth credentials.
 
-        # Attempt to load cached credentials
+        This method attempts to:
-        if os.path.exists(self.token_path):
+        1. Load cached credentials from the configured credential store
-            try:
+        2. Refresh expired credentials when possible
-                with open(self.token_path, "rb") as fh:
+        3. Perform an interactive OAuth login as a fallback
-                    creds = pickle.load(fh)
+        4. Persist valid credentials for future use
-            except Exception:
+
-                creds = None
+        Returns:
+            A ``google.oauth2.credentials.Credentials`` instance suitable
+            for use with Google API clients.
+
+        Raises:
+            MailIntakeAuthError: If credentials cannot be loaded, refreshed,
+                or obtained via interactive authentication.
+        """
+        creds = self.store.load()
 
         # Validate / refresh credentials
         if not creds or not creds.valid:
@@ -49,6 +93,7 @@ class MailIntakeGoogleAuth(MailIntakeAuthProvider):
                 try:
                     creds.refresh(Request())
                 except google.auth.exceptions.RefreshError:
+                    self.store.clear()
                     creds = None
 
             # Interactive login if refresh failed or creds missing
@@ -69,13 +114,12 @@ class MailIntakeGoogleAuth(MailIntakeAuthProvider):
                         "Failed to complete Google OAuth flow"
                     ) from exc
 
-            # Persist refreshed / new credentials
+            # Persist refreshed or newly obtained credentials
             try:
-                with open(self.token_path, "wb") as fh:
+                self.store.save(creds)
-                    pickle.dump(creds, fh)
             except Exception as exc:
                 raise MailIntakeAuthError(
-                    f"Failed to write token file: {self.token_path}"
+                    "Failed to persist Google OAuth credentials"
                 ) from exc
 
         return creds

mail_intake/config.py

@@ -1,3 +1,14 @@
+"""
+Global configuration models for Mail Intake.
+
+This module defines the **top-level configuration object** used to control
+mail ingestion behavior across adapters, authentication providers, and
+ingestion workflows.
+
+Configuration is intentionally explicit, immutable, and free of implicit
+environment reads to ensure predictability and testability.
+"""
+
 from dataclasses import dataclass
 from typing import Optional
 
@@ -9,12 +20,26 @@ class MailIntakeConfig:
 
     This configuration is intentionally explicit and immutable.
     No implicit environment reads or global state.
+
+    Design principles:
+    - Immutable once constructed
+    - Explicit configuration over implicit defaults
+    - No direct environment or filesystem access
+
+    This model is safe to pass across layers and suitable for serialization.
     """
 
     provider: str = "gmail"
-    user_id: str = "me"
+    """Identifier of the mail provider to use (e.g., ``"gmail"``)."""
-    readonly: bool = True
+
+    user_id: str = "me"
+    """Provider-specific user identifier. Defaults to the authenticated user."""
+
+    readonly: bool = True
+    """Whether ingestion should operate in read-only mode."""
 
-    # Provider-specific paths (optional at this layer)
     credentials_path: Optional[str] = None
+    """Optional path to provider credentials configuration."""
+
     token_path: Optional[str] = None
+    """Optional path to persisted authentication tokens."""

mail_intake/credentials/__init__.py

@@ -0,0 +1,29 @@
+"""
+Credential persistence interfaces and implementations for Mail Intake.
+
+This package defines the abstractions and concrete implementations used
+to persist authentication credentials across Mail Intake components.
+
+The credential persistence layer is intentionally decoupled from
+authentication logic. Authentication providers are responsible for
+credential acquisition, validation, and refresh, while implementations
+within this package are responsible solely for storage and retrieval.
+
+The package provides:
+- A generic ``CredentialStore`` abstraction defining the persistence contract
+- Local filesystem–based storage for development and single-node use
+- Distributed, Redis-backed storage for production and scaled deployments
+
+Credential lifecycle management, interpretation, and security policy
+decisions remain the responsibility of authentication providers.
+"""
+
+from mail_intake.credentials.store import CredentialStore
+from mail_intake.credentials.pickle import PickleCredentialStore
+from mail_intake.credentials.redis import RedisCredentialStore
+
+__all__ = [
+    "CredentialStore",
+    "PickleCredentialStore",
+    "RedisCredentialStore",
+]

mail_intake/credentials/pickle.py

@@ -0,0 +1,96 @@
+"""
+Local filesystem–based credential persistence for Mail Intake.
+
+This module provides a file-backed implementation of the
+``CredentialStore`` abstraction using Python's ``pickle`` module.
+
+The pickle-based credential store is intended for local development,
+single-node deployments, and controlled environments where credentials
+do not need to be shared across processes or machines.
+
+Due to the security and portability risks associated with pickle-based
+serialization, this implementation is not suitable for distributed or
+untrusted environments.
+"""
+
+import pickle
+from typing import Optional, TypeVar
+
+from mail_intake.credentials.store import CredentialStore
+
+T = TypeVar("T")
+
+
+class PickleCredentialStore(CredentialStore[T]):
+    """
+    Filesystem-backed credential store using pickle serialization.
+
+    This store persists credentials as a pickled object on the local
+    filesystem. It is a simple implementation intended primarily for
+    development, testing, and single-process execution contexts.
+
+    This implementation:
+    - Stores credentials on the local filesystem
+    - Uses pickle for serialization and deserialization
+    - Does not provide encryption, locking, or concurrency guarantees
+
+    Credential lifecycle management, validation, and refresh logic are
+    explicitly out of scope for this class.
+    """
+
+    def __init__(self, path: str):
+        """
+        Initialize a pickle-backed credential store.
+
+        Args:
+            path:
+                Filesystem path where credentials will be stored.
+                The file will be created or overwritten as needed.
+        """
+        self.path = path
+
+    def load(self) -> Optional[T]:
+        """
+        Load credentials from the local filesystem.
+
+        If the credential file does not exist or cannot be successfully
+        deserialized, this method returns ``None``.
+
+        The store does not attempt to validate or interpret the returned
+        credentials.
+
+        Returns:
+            An instance of type ``T`` if credentials are present and
+            successfully deserialized; otherwise ``None``.
+        """
+        try:
+            with open(self.path, "rb") as fh:
+                return pickle.load(fh)
+        except Exception:
+            return None
+
+    def save(self, credentials: T) -> None:
+        """
+        Persist credentials to the local filesystem.
+
+        Any previously stored credentials at the configured path are
+        overwritten.
+
+        Args:
+            credentials:
+                The credential object to persist.
+        """
+        with open(self.path, "wb") as fh:
+            pickle.dump(credentials, fh)
+
+    def clear(self) -> None:
+        """
+        Remove persisted credentials from the local filesystem.
+
+        This method deletes the credential file if it exists and should
+        be treated as an idempotent operation.
+        """
+        import os
+
+        if os.path.exists(self.path):
+            os.remove(self.path)

mail_intake/credentials/redis.py

@@ -0,0 +1,142 @@
+"""
+Redis-backed credential persistence for Mail Intake.
+
+This module provides a Redis-based implementation of the
+``CredentialStore`` abstraction, enabling credential persistence
+across distributed and horizontally scaled deployments.
+
+The Redis credential store is designed for environments where
+authentication credentials must be shared safely across multiple
+processes, containers, or nodes, such as container orchestration
+platforms and microservice architectures.
+
+Key characteristics:
+- Distributed-safe, shared storage using Redis
+- Explicit, caller-defined serialization and deserialization
+- No reliance on unsafe mechanisms such as pickle
+- Optional time-to-live (TTL) support for automatic credential expiry
+
+This module is responsible solely for persistence concerns.
+Credential validation, refresh, rotation, and acquisition remain the
+responsibility of authentication provider implementations.
+"""
+
+
+from typing import Optional, TypeVar, Callable
+
+from mail_intake.credentials.store import CredentialStore
+
+T = TypeVar("T")
+
+
+class RedisCredentialStore(CredentialStore[T]):
+    """
+    Redis-backed implementation of ``CredentialStore``.
+
+    This store persists credentials in Redis and is suitable for
+    distributed and horizontally scaled deployments where credentials
+    must be shared across multiple processes or nodes.
+
+    The store is intentionally generic and delegates all serialization
+    concerns to caller-provided functions. This avoids unsafe mechanisms
+    such as pickle and allows credential formats to be explicitly
+    controlled and audited.
+
+    This class is responsible only for persistence and retrieval.
+    It does not interpret, validate, refresh, or otherwise manage
+    the lifecycle of the credentials being stored.
+    """
+
+    def __init__(
+        self,
+        redis_client,
+        key: str,
+        serialize: Callable[[T], bytes],
+        deserialize: Callable[[bytes], T],
+        ttl_seconds: Optional[int] = None,
+    ):
+        """
+        Initialize a Redis-backed credential store.
+
+        Args:
+            redis_client:
+                An initialized Redis client instance (for example,
+                ``redis.Redis`` or a compatible interface) used to
+                communicate with the Redis server.
+
+            key:
+                The Redis key under which credentials are stored.
+                Callers are responsible for applying appropriate
+                namespacing to avoid collisions.
+
+            serialize:
+                A callable that converts a credential object of type
+                ``T`` into a ``bytes`` representation suitable for
+                storage in Redis.
+
+            deserialize:
+                A callable that converts a ``bytes`` payload retrieved
+                from Redis back into a credential object of type ``T``.
+
+            ttl_seconds:
+                Optional time-to-live (TTL) for the stored credentials,
+                expressed in seconds. When provided, Redis will
+                automatically expire the stored credentials after the
+                specified duration. If ``None``, credentials are stored
+                without an expiration.
+        """
+        self.redis = redis_client
+        self.key = key
+        self.serialize = serialize
+        self.deserialize = deserialize
+        self.ttl_seconds = ttl_seconds
+
+    def load(self) -> Optional[T]:
+        """
+        Load credentials from Redis.
+
+        If no value exists for the configured key, or if the stored
+        payload cannot be successfully deserialized, this method
+        returns ``None``.
+
+        The store does not attempt to validate the returned credentials
+        or determine whether they are expired or otherwise usable.
+
+        Returns:
+            An instance of type ``T`` if credentials are present and
+            successfully deserialized; otherwise ``None``.
+        """
+        raw = self.redis.get(self.key)
+        if not raw:
+            return None
+        try:
+            return self.deserialize(raw)
+        except Exception:
+            return None
+
+    def save(self, credentials: T) -> None:
+        """
+        Persist credentials to Redis.
+
+        Any previously stored credentials under the same key are
+        overwritten. If a TTL is configured, the credentials will
+        expire automatically after the specified duration.
+
+        Args:
+            credentials:
+                The credential object to persist.
+        """
+        payload = self.serialize(credentials)
+        if self.ttl_seconds:
+            self.redis.setex(self.key, self.ttl_seconds, payload)
+        else:
+            self.redis.set(self.key, payload)
+
+    def clear(self) -> None:
+        """
+        Remove stored credentials from Redis.
+
+        This operation deletes the configured Redis key if it exists.
+        Implementations should treat this method as idempotent.
+        """
+        self.redis.delete(self.key)

mail_intake/credentials/store.py

@@ -0,0 +1,96 @@
+"""
+Credential persistence abstractions for Mail Intake.
+
+This module defines the generic persistence contract used to store and
+retrieve authentication credentials across Mail Intake components.
+
+The ``CredentialStore`` abstraction establishes a strict separation
+between credential *lifecycle management* and credential *storage*.
+Authentication providers are responsible for acquiring, validating,
+refreshing, and revoking credentials, while concrete store
+implementations are responsible solely for persistence concerns.
+
+By remaining agnostic to credential structure, serialization format,
+and storage backend, this module enables multiple persistence
+strategies—such as local files, in-memory caches, distributed stores,
+or secrets managers—without coupling authentication logic to any
+specific storage mechanism.
+"""
+
+
+from abc import ABC, abstractmethod
+from typing import Generic, Optional, TypeVar
+
+T = TypeVar("T")
+
+
+class CredentialStore(ABC, Generic[T]):
+    """
+    Abstract base class defining a generic persistence interface for
+    authentication credentials.
+
+    This interface separates *credential lifecycle management* from
+    *credential storage mechanics*. Implementations are responsible
+    only for persistence concerns, while authentication providers
+    retain full control over credential creation, validation, refresh,
+    and revocation logic.
+
+    The store is intentionally agnostic to:
+    - The concrete credential type being stored
+    - The serialization format used to persist credentials
+    - The underlying storage backend or durability guarantees
+
+    Type Parameters:
+        T:
+            The concrete credential type managed by the store. This may
+            represent OAuth credentials, API tokens, session objects,
+            or any other authentication material.
+    """
+
+    @abstractmethod
+    def load(self) -> Optional[T]:
+        """
+        Load previously persisted credentials.
+
+        Implementations should return ``None`` when no credentials are
+        present or when stored credentials cannot be successfully
+        decoded or deserialized.
+
+        The store must not attempt to validate, refresh, or otherwise
+        interpret the returned credentials.
+
+        Returns:
+            An instance of type ``T`` if credentials are available and
+            loadable; otherwise ``None``.
+        """
+
+    @abstractmethod
+    def save(self, credentials: T) -> None:
+        """
+        Persist credentials to the underlying storage backend.
+
+        This method is invoked when credentials are newly obtained or
+        have been refreshed and are known to be valid at the time of
+        persistence.
+
+        Implementations are responsible for:
+        - Ensuring durability appropriate to the deployment context
+        - Applying encryption or access controls where required
+        - Overwriting any previously stored credentials
+
+        Args:
+            credentials:
+                The credential object to persist.
+        """
+
+    @abstractmethod
+    def clear(self) -> None:
+        """
+        Remove any persisted credentials from the store.
+
+        This method is called when credentials are known to be invalid,
+        revoked, corrupted, or otherwise unusable, and must ensure that
+        no stale authentication material remains accessible.
+
+        Implementations should treat this operation as idempotent.
+        """

mail_intake/exceptions.py

@@ -1,19 +1,49 @@
+"""
+Exception hierarchy for Mail Intake.
+
+This module defines the **canonical exception types** used throughout the
+Mail Intake library.
+
+All library-raised errors derive from `MailIntakeError`. Consumers are
+encouraged to catch this base type (or specific subclasses) rather than
+provider-specific or third-party exceptions.
+"""
+
+
 class MailIntakeError(Exception):
     """
-    Base exception for all mail-intake errors.
+    Base exception for all Mail Intake errors.
 
-    Users of the library should catch this type (or subclasses)
+    This is the root of the Mail Intake exception hierarchy.
-    instead of provider-specific or third-party exceptions.
+    All errors raised by the library must derive from this class.
+
+    Consumers should generally catch this type when handling
+    library-level failures.
     """
 
 
 class MailIntakeAuthError(MailIntakeError):
-    """Authentication and credential-related failures."""
+    """
+    Authentication and credential-related failures.
+
+    Raised when authentication providers are unable to acquire,
+    refresh, or persist valid credentials.
+    """
 
 
 class MailIntakeAdapterError(MailIntakeError):
-    """Errors raised by mail provider adapters."""
+    """
+    Errors raised by mail provider adapters.
+
+    Raised when a provider adapter encounters API errors,
+    transport failures, or invalid provider responses.
+    """
 
 
 class MailIntakeParsingError(MailIntakeError):
-    """Errors encountered while parsing message content."""
+    """
+    Errors encountered while parsing message content.
+
+    Raised when raw provider payloads cannot be interpreted
+    or normalized into internal domain models.
+    """

mail_intake/ingestion/__init__.py

@@ -0,0 +1,24 @@
+"""
+Mail ingestion orchestration for Mail Intake.
+
+This package contains **high-level ingestion components** responsible for
+coordinating mail retrieval, parsing, normalization, and model construction.
+
+It represents the **top of the ingestion pipeline** and is intended to be the
+primary interaction surface for library consumers.
+
+Components in this package:
+- Are provider-agnostic
+- Depend only on adapter and parser contracts
+- Contain no provider-specific API logic
+- Expose read-only ingestion workflows
+
+Consumers are expected to construct a mail adapter and pass it to the
+ingestion layer to begin processing messages and threads.
+"""
+
+from .reader import MailIntakeReader
+
+__all__ = [
+    "MailIntakeReader",
+]

mail_intake/ingestion/reader.py

@@ -1,3 +1,18 @@
+"""
+High-level mail ingestion orchestration for Mail Intake.
+
+This module provides the primary, provider-agnostic entry point for
+reading and processing mail data.
+
+It coordinates:
+- Mail adapter access
+- Message and thread iteration
+- Header and body parsing
+- Normalization and model construction
+
+No provider-specific logic or API semantics are permitted in this layer.
+"""
+
 from datetime import datetime
 from typing import Iterator, Dict, Any
 
@@ -14,22 +29,43 @@ class MailIntakeReader:
     """
     High-level read-only ingestion interface.
 
-    This is the primary entry point users should interact with.
+    This class is the **primary entry point** for consumers of the Mail
-    It orchestrates:
+    Intake library.
-    - adapter calls
-    - parsing
-    - normalization
-    - model construction
 
-    No provider-specific logic exists here.
+    It orchestrates the full ingestion pipeline:
+    - Querying the adapter for message references
+    - Fetching raw provider messages
+    - Parsing and normalizing message data
+    - Constructing domain models
+
+    This class is intentionally:
+    - Provider-agnostic
+    - Stateless beyond iteration scope
+    - Read-only
     """
 
     def __init__(self, adapter: MailIntakeAdapter):
+        """
+        Initialize the mail reader.
+
+        Args:
+            adapter: Mail adapter implementation used to retrieve raw
+                messages and threads from a mail provider.
+        """
         self._adapter = adapter
 
     def iter_messages(self, query: str) -> Iterator[MailIntakeMessage]:
         """
         Iterate over parsed messages matching a provider query.
+
+        Args:
+            query: Provider-specific query string used to filter messages.
+
+        Yields:
+            Fully parsed and normalized `MailIntakeMessage` instances.
+
+        Raises:
+            MailIntakeParsingError: If a message cannot be parsed.
         """
         for ref in self._adapter.iter_message_refs(query):
             raw = self._adapter.fetch_message(ref["message_id"])
@@ -39,7 +75,17 @@ class MailIntakeReader:
         """
         Iterate over threads constructed from messages matching a query.
 
-        Messages are grouped by thread_id and yielded as complete threads.
+        Messages are grouped by `thread_id` and yielded as complete thread
+        objects containing all associated messages.
+
+        Args:
+            query: Provider-specific query string used to filter messages.
+
+        Returns:
+            An iterator of `MailIntakeThread` instances.
+
+        Raises:
+            MailIntakeParsingError: If a message cannot be parsed.
         """
         threads: Dict[str, MailIntakeThread] = {}
 
@@ -61,7 +107,17 @@ class MailIntakeReader:
 
     def _parse_message(self, raw_message: Dict[str, Any]) -> MailIntakeMessage:
         """
-        Parse a raw provider message into a MailIntakeMessage.
+        Parse a raw provider message into a `MailIntakeMessage`.
+
+        Args:
+            raw_message: Provider-native message payload.
+
+        Returns:
+            A fully populated `MailIntakeMessage` instance.
+
+        Raises:
+            MailIntakeParsingError: If the message payload is missing required
+                fields or cannot be parsed.
         """
         try:
             message_id = raw_message["id"]

mail_intake/models/__init__.py

@@ -0,0 +1,22 @@
+"""
+Domain models for Mail Intake.
+
+This package defines the **canonical, provider-agnostic data models**
+used throughout the Mail Intake ingestion pipeline.
+
+Models in this package:
+- Represent fully parsed and normalized mail data
+- Are safe to persist, serialize, and index
+- Contain no provider-specific payloads or API semantics
+- Serve as stable inputs for downstream processing and analysis
+
+These models form the core internal data contract of the library.
+"""
+
+from .message import MailIntakeMessage
+from .thread import MailIntakeThread
+
+__all__ = [
+    "MailIntakeMessage",
+    "MailIntakeThread",
+]

mail_intake/models/message.py

@@ -1,3 +1,14 @@
+"""
+Message domain models for Mail Intake.
+
+This module defines the **canonical, provider-agnostic representation**
+of an individual email message as used internally by the Mail Intake
+ingestion pipeline.
+
+Models in this module are safe to persist and must not contain any
+provider-specific fields or semantics.
+"""
+
 from dataclasses import dataclass
 from datetime import datetime
 from typing import Optional, Dict
@@ -8,19 +19,37 @@ class MailIntakeMessage:
     """
     Canonical internal representation of a single email message.
 
-    This model is provider-agnostic and safe to persist.
+    This model represents a fully parsed and normalized email message.
-    No Gmail-specific fields should appear here.
+    It is intentionally provider-agnostic and suitable for persistence,
+    indexing, and downstream processing.
+
+    No provider-specific identifiers, payloads, or API semantics
+    should appear in this model.
     """
 
     message_id: str
+    """Provider-specific message identifier."""
+
     thread_id: str
+    """Provider-specific thread identifier to which this message belongs."""
+
     timestamp: datetime
+    """Message timestamp as a timezone-naive UTC datetime."""
 
     from_email: str
+    """Sender email address."""
+
     from_name: Optional[str]
+    """Optional human-readable sender name."""
 
     subject: str
+    """Raw subject line of the message."""
+
     body_text: str
+    """Extracted plain-text body content of the message."""
+
     snippet: str
+    """Short provider-supplied preview snippet of the message."""
 
     raw_headers: Dict[str, str]
+    """Normalized mapping of message headers (header name → value)."""

mail_intake/models/thread.py

@@ -1,3 +1,13 @@
+"""
+Thread domain models for Mail Intake.
+
+This module defines the **canonical, provider-agnostic representation**
+of an email thread as used internally by the Mail Intake ingestion pipeline.
+
+Threads group related messages and serve as the primary unit of reasoning
+for higher-level correspondence workflows.
+"""
+
 from dataclasses import dataclass, field
 from datetime import datetime
 from typing import List, Set
@@ -10,21 +20,40 @@ class MailIntakeThread:
     """
     Canonical internal representation of an email thread.
 
-    Threads are the primary unit of reasoning for correspondence
+    A thread groups multiple related messages under a single subject
-    workflows (job applications, interviews, follow-ups, etc.).
+    and participant set. It is designed to support reasoning over
+    conversational context such as job applications, interviews,
+    follow-ups, and ongoing discussions.
+
+    This model is provider-agnostic and safe to persist.
     """
 
     thread_id: str
+    """Provider-specific thread identifier."""
+
     normalized_subject: str
+    """Normalized subject line used to group related messages."""
 
     participants: Set[str] = field(default_factory=set)
+    """Set of unique participant email addresses observed in the thread."""
+
     messages: List[MailIntakeMessage] = field(default_factory=list)
+    """Ordered list of messages belonging to this thread."""
 
     last_activity_at: datetime | None = None
+    """Timestamp of the most recent message in the thread."""
 
     def add_message(self, message: MailIntakeMessage) -> None:
         """
         Add a message to the thread and update derived fields.
+
+        This method:
+        - Appends the message to the thread
+        - Tracks unique participants
+        - Updates the last activity timestamp
+
+        Args:
+            message: Parsed mail message to add to the thread.
         """
         self.messages.append(message)
 

mail_intake/parsers/__init__.py

@@ -0,0 +1,30 @@
+"""
+Message parsing utilities for Mail Intake.
+
+This package contains **provider-aware but adapter-agnostic parsing helpers**
+used to extract and normalize structured information from raw mail payloads.
+
+Parsers in this package are responsible for:
+- Interpreting provider-native message structures
+- Extracting meaningful fields such as headers, body text, and subjects
+- Normalizing data into consistent internal representations
+
+This package does not:
+- Perform network or IO operations
+- Contain provider API logic
+- Construct domain models directly
+
+Parsing functions are designed to be composable and are orchestrated by the
+ingestion layer.
+"""
+
+from .body import extract_body
+from .headers import parse_headers, extract_sender
+from .subject import normalize_subject
+
+__all__ = [
+    "extract_body",
+    "parse_headers",
+    "extract_sender",
+    "normalize_subject",
+]

mail_intake/parsers/body.py

@@ -1,3 +1,13 @@
+"""
+Message body extraction utilities for Mail Intake.
+
+This module contains helper functions for extracting a best-effort
+plain-text body from provider-native message payloads.
+
+The logic is intentionally tolerant of malformed or partial data and
+prefers human-readable text over fidelity to original formatting.
+"""
+
 import base64
 from typing import Dict, Any, Optional
 
@@ -9,6 +19,18 @@ from mail_intake.exceptions import MailIntakeParsingError
 def _decode_base64(data: str) -> str:
     """
     Decode Gmail URL-safe base64 payload into UTF-8 text.
+
+    Gmail message bodies are encoded using URL-safe base64, which may
+    omit padding and use non-standard characters.
+
+    Args:
+        data: URL-safe base64-encoded string.
+
+    Returns:
+        Decoded UTF-8 text with replacement for invalid characters.
+
+    Raises:
+        MailIntakeParsingError: If decoding fails.
     """
     try:
         padded = data.replace("-", "+").replace("_", "/")
@@ -21,6 +43,16 @@ def _decode_base64(data: str) -> str:
 def _extract_from_part(part: Dict[str, Any]) -> Optional[str]:
     """
     Extract text content from a single MIME part.
+
+    Supports:
+    - text/plain
+    - text/html (converted to plain text)
+
+    Args:
+        part: MIME part dictionary from a provider payload.
+
+    Returns:
+        Extracted plain-text content, or None if unsupported or empty.
     """
     mime_type = part.get("mimeType")
     body = part.get("body", {})
@@ -49,7 +81,14 @@ def extract_body(payload: Dict[str, Any]) -> str:
     Priority:
     1. text/plain
     2. text/html (stripped to text)
-    3. empty string (if nothing usable found)
+    3. Single-part body
+    4. empty string (if nothing usable found)
+
+    Args:
+        payload: Provider-native message payload dictionary.
+
+    Returns:
+        Extracted plain-text message body.
     """
     if not payload:
         return ""

mail_intake/parsers/headers.py

@@ -1,3 +1,13 @@
+"""
+Message header parsing utilities for Mail Intake.
+
+This module provides helper functions for normalizing and extracting
+useful information from provider-native message headers.
+
+The functions here are intentionally simple and tolerant of malformed
+or incomplete header data.
+"""
+
 from typing import Dict, List, Tuple, Optional
 
 
@@ -5,18 +15,28 @@ def parse_headers(raw_headers: List[Dict[str, str]]) -> Dict[str, str]:
     """
     Convert a list of Gmail-style headers into a normalized dict.
 
+    Provider payloads (such as Gmail) typically represent headers as a list
+    of name/value mappings. This function normalizes them into a
+    case-insensitive dictionary keyed by lowercase header names.
+
+    Args:
+        raw_headers: List of header dictionaries, each containing
+            ``name`` and ``value`` keys.
+
+    Returns:
+        Dictionary mapping lowercase header names to stripped values.
+
+    Example:
         Input:
             [
                 {"name": "From", "value": "John Doe <john@example.com>"},
                 {"name": "Subject", "value": "Re: Interview Update"},
-        ...
             ]
 
         Output:
             {
-        "from": "...",
+                "from": "John Doe <john@example.com>",
-        "subject": "...",
+                "subject": "Re: Interview Update",
-        ...
             }
     """
     headers: Dict[str, str] = {}
@@ -37,18 +57,27 @@ def extract_sender(headers: Dict[str, str]) -> Tuple[str, Optional[str]]:
     """
     Extract sender email and optional display name from headers.
 
-    Returns:
+    This function parses the ``From`` header and attempts to extract:
-        (email, name)
+    - Sender email address
+    - Optional human-readable display name
 
-    If name cannot be determined, name will be None.
+    Args:
+        headers: Normalized header dictionary as returned by
+            :func:`parse_headers`.
+
+    Returns:
+        A tuple ``(email, name)`` where:
+        - ``email`` is the sender email address
+        - ``name`` is the display name, or ``None`` if unavailable
+
+    Examples:
+        ``"John Doe <john@example.com>"`` → ``("john@example.com", "John Doe")``
+        ``"john@example.com"`` → ``("john@example.com", None)``
     """
     from_header = headers.get("from")
     if not from_header:
         return "", None
 
-    # Common forms:
-    #   Name <email@domain>
-    #   email@domain
     if "<" in from_header and ">" in from_header:
         name_part, email_part = from_header.split("<", 1)
         email = email_part.rstrip(">").strip()

mail_intake/parsers/subject.py

@@ -1,7 +1,18 @@
+"""
+Subject line normalization utilities for Mail Intake.
+
+This module provides helper functions for normalizing email subject lines
+to enable reliable thread-level comparison and grouping.
+
+Normalization is intentionally conservative to avoid altering semantic
+meaning while removing common reply and forward prefixes.
+"""
+
 import re
 
 
 _PREFIX_RE = re.compile(r"^(re|fw|fwd)\s*:\s*", re.IGNORECASE)
+"""Regular expression matching common reply/forward subject prefixes."""
 
 
 def normalize_subject(subject: str) -> str:
@@ -9,11 +20,19 @@ def normalize_subject(subject: str) -> str:
     Normalize an email subject for thread-level comparison.
 
     Operations:
-    - Strip common prefixes (Re:, Fwd:, FW:)
+    - Strips common prefixes such as ``Re:``, ``Fwd:``, and ``FW:``
-    - Collapse whitespace
+    - Repeats prefix stripping to handle stacked prefixes
-    - Preserve original casing (no lowercasing)
+    - Collapses excessive whitespace
+    - Preserves original casing (no lowercasing)
 
-    This function is intentionally conservative.
+    This function is intentionally conservative and avoids aggressive
+    transformations that could alter the semantic meaning of the subject.
+
+    Args:
+        subject: Raw subject line from a message header.
+
+    Returns:
+        Normalized subject string suitable for thread grouping.
     """
     if not subject:
         return ""

manage_docs.py

@@ -0,0 +1,159 @@
+"""
+MkDocs documentation management CLI.
+
+This script provides a proper CLI interface to:
+- Generate MkDocs Markdown files with mkdocstrings directives
+- Build the documentation site
+- Serve the documentation site locally
+
+All operations are performed by calling MkDocs as a Python library
+(no shell command invocation).
+
+Requirements:
+- mkdocs
+- mkdocs-material
+- mkdocstrings[python]
+
+Usage:
+    python manage_docs.py generate
+    python manage_docs.py build
+    python manage_docs.py serve
+
+Optional flags:
+    --docs-dir PATH        Path to docs directory (default: ./docs)
+    --package-root NAME    Root Python package name (default: mail_intake)
+"""
+
+from __future__ import annotations
+
+import argparse
+from pathlib import Path
+
+from mkdocs.commands import build as mkdocs_build
+from mkdocs.commands import serve as mkdocs_serve
+from mkdocs.config import load_config
+
+
+PROJECT_ROOT = Path(__file__).resolve().parent
+DEFAULT_DOCS_DIR = PROJECT_ROOT / "docs"
+DEFAULT_PACKAGE_ROOT = "mail_intake"
+MKDOCS_YML = PROJECT_ROOT / "mkdocs.yml"
+
+
+def generate_docs_from_nav(
+    project_root: Path,
+    docs_root: Path,
+    package_root: str,
+) -> None:
+    """
+    Create and populate MkDocs Markdown files with mkdocstrings directives.
+
+    This function:
+    - Walks the Python package structure
+    - Mirrors it under the docs directory
+    - Creates missing .md files
+    - Creates index.md for packages (__init__.py)
+    - Overwrites content with ::: package.module
+
+    Examples:
+        mail_intake/__init__.py           -> docs/mail_intake/index.md
+        mail_intake/config.py             -> docs/mail_intake/config.md
+        mail_intake/adapters/__init__.py  -> docs/mail_intake/adapters/index.md
+        mail_intake/adapters/base.py      -> docs/mail_intake/adapters/base.md
+    """
+
+    package_dir = project_root / package_root
+    if not package_dir.exists():
+        raise FileNotFoundError(f"Package not found: {package_dir}")
+
+    docs_root.mkdir(parents=True, exist_ok=True)
+
+    for py_file in package_dir.rglob("*.py"):
+        rel = py_file.relative_to(project_root)
+
+        if py_file.name == "__init__.py":
+            # Package → index.md
+            module_path = ".".join(rel.parent.parts)
+            md_path = docs_root / rel.parent / "index.md"
+            title = rel.parent.name.replace("_", " ").title()
+        else:
+            # Regular module → <module>.md
+            module_path = ".".join(rel.with_suffix("").parts)
+            md_path = docs_root / rel.with_suffix(".md")
+            title = md_path.stem.replace("_", " ").title()
+
+        md_path.parent.mkdir(parents=True, exist_ok=True)
+
+        content = f"""# {title}
+
+::: {module_path}
+"""
+
+        md_path.write_text(content, encoding="utf-8")
+
+
+def load_mkdocs_config():
+    if not MKDOCS_YML.exists():
+        raise FileNotFoundError("mkdocs.yml not found at project root")
+    return load_config(str(MKDOCS_YML))
+
+
+def cmd_generate(args: argparse.Namespace) -> None:
+    generate_docs_from_nav(
+        project_root=PROJECT_ROOT,
+        docs_root=args.docs_dir,
+        package_root=args.package_root,
+    )
+
+
+def cmd_build(_: argparse.Namespace) -> None:
+    config = load_mkdocs_config()
+    mkdocs_build.build(config)
+
+
+def cmd_serve(_: argparse.Namespace) -> None:
+    mkdocs_serve.serve(
+        config_file=str(MKDOCS_YML)
+    )
+
+def main() -> None:
+    parser = argparse.ArgumentParser(
+        prog="manage_docs.py",
+        description="Manage MkDocs documentation for the project",
+    )
+
+    parser.add_argument(
+        "--docs-dir",
+        type=Path,
+        default=DEFAULT_DOCS_DIR,
+        help="Path to the docs directory",
+    )
+    parser.add_argument(
+        "--package-root",
+        default=DEFAULT_PACKAGE_ROOT,
+        help="Root Python package name",
+    )
+
+    subparsers = parser.add_subparsers(dest="command", required=True)
+
+    subparsers.add_parser(
+        "generate",
+        help="Generate Markdown files with mkdocstrings directives",
+    ).set_defaults(func=cmd_generate)
+
+    subparsers.add_parser(
+        "build",
+        help="Build the MkDocs site",
+    ).set_defaults(func=cmd_build)
+
+    subparsers.add_parser(
+        "serve",
+        help="Serve the MkDocs site locally",
+    ).set_defaults(func=cmd_serve)
+
+    args = parser.parse_args()
+    args.func(args)
+
+
+if __name__ == "__main__":
+    main()

mkdocs.yml

@@ -0,0 +1,66 @@
+site_name: Aetoskia Mail Intake
+site_description: Format-agnostic document reading, parsing, and scraping framework
+
+theme:
+  name: material
+  palette:
+    - scheme: slate
+      primary: deep purple
+      accent: cyan
+  font:
+    text: Inter
+    code: JetBrains Mono
+  features:
+    - navigation.tabs
+    - navigation.expand
+    - navigation.top
+    - navigation.instant
+    - content.code.copy
+    - content.code.annotate
+
+plugins:
+  - search
+  - mkdocstrings:
+      handlers:
+        python:
+          paths: ["."]
+          options:
+            docstring_style: google
+            show_source: false
+            show_signature_annotations: true
+            separate_signature: true
+            merge_init_into_class: true
+            inherited_members: true
+            annotations_path: brief
+            show_root_heading: true
+            group_by_category: true
+
+nav:
+  - Home: mail_intake/index.md
+
+  - Adapters:
+      - Base Adapter: mail_intake/adapters/base.md
+      - Gmail Adapter: mail_intake/adapters/gmail.md
+
+  - Auth:
+      - Base Auth: mail_intake/auth/base.md
+      - Google Auth: mail_intake/auth/google.md
+
+  - Credentials Store:
+      - Store: mail_intake/credentials/store.md
+      - Pickle: mail_intake/credentials/pickle.md
+      - Redis: mail_intake/credentials/redis.md
+
+  - Mail Reader: mail_intake/ingestion/reader.md
+
+  - Models:
+      - Message: mail_intake/models/message.md
+      - Thread: mail_intake/models/thread.md
+
+  - Parsers:
+      - Body: mail_intake/parsers/body.md
+      - Headers: mail_intake/parsers/headers.md
+      - Subject: mail_intake/parsers/subject.md
+
+  - Config: mail_intake/config.md
+  - Exceptions: mail_intake/exceptions.md

pyproject.toml

@@ -5,7 +5,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "mail-intake"
-version = "0.0.1"
+version = "0.0.2"
 description = "Structured mail ingestion and correspondence parsing with provider adapters (Gmail-first)."
 readme = "README.md"
 requires-python = ">=3.10"
@@ -77,11 +77,7 @@ Versions = "https://git.aetoskia.com/aetos/mail-intake/tags"
 
 
 [tool.setuptools]
-package-dir = { "" = "src" }
+packages = { find = { include = ["mail_intake*"] } }
-
-[tool.setuptools.packages.find]
-where = ["src"]
-include = ["mail_intake*"]
 
 
 [tool.ruff]

requirements.txt

@@ -1,10 +1,17 @@
 beautifulsoup4==4.12.0
+google-api-python-client==2.187.0
+google-auth-oauthlib==1.2.3
+types-beautifulsoup4
 
+# Test Packages
 pytest==7.4.0
 pytest-asyncio==0.21.0
 pytest-cov==4.1.0
 
-types-beautifulsoup4
+# Doc Packages
-
+mkdocs==1.6.1
-# Optional, useful locally
+mkdocs-material==9.6.23
-ipython
+neoteroi-mkdocs==1.1.3
+pymdown-extensions==10.16.1
+mkdocstrings==1.0.0
+mkdocstrings-python==2.0.1

tests/unit/test_models.py

@@ -1,7 +1,7 @@
 from datetime import datetime, timedelta
 
-from mail_intake.models.message import MailIntakeMessage
+from mail_intake.models import MailIntakeMessage
-from mail_intake.models.thread import MailIntakeThread
+from mail_intake.models import MailIntakeThread
 
 
 def test_message_is_immutable():

tests/unit/test_parsers.py

@@ -1,8 +1,8 @@
 import base64
 
-from mail_intake.parsers.subject import normalize_subject
+from mail_intake.parsers import normalize_subject
-from mail_intake.parsers.headers import parse_headers, extract_sender
+from mail_intake.parsers import parse_headers, extract_sender
-from mail_intake.parsers.body import extract_body
+from mail_intake.parsers import extract_body
 
 
 def _b64(text: str) -> str: