Add configurable LLM provider adapters for email classification

app/classifier.py

@@ -0,0 +1,91 @@
+from __future__ import annotations
+
+import json
+from typing import Any
+
+from app.config import get_request_settings
+from app.llm_adapters import build_adapter, coerce_json_text
+from app.models import ClassificationResult, ClassifyRequest, EmailData
+
+VALID_CATEGORIES = {
+    "action_required",
+    "question",
+    "fyi",
+    "newsletter",
+    "promotional",
+    "automated",
+    "alert",
+    "uncategorized",
+}
+VALID_PRIORITIES = {"high", "medium", "low"}
+
+
+async def classify_email(request: ClassifyRequest) -> ClassificationResult:
+    clean_email = _clean_email(request.email_data)
+    settings = get_request_settings(
+        provider=request.provider,
+        model=request.model,
+        base_url=request.base_url,
+        api_key=request.api_key,
+        temperature=request.temperature,
+    )
+    adapter = build_adapter(settings)
+
+    attempts = 0
+    while attempts < settings.max_retries:
+        raw_response = await adapter.classify(clean_email)
+        try:
+            payload = json.loads(coerce_json_text(raw_response))
+            result = _normalize_result(payload)
+            if result.needs_action and not result.task_description:
+                attempts += 1
+                continue
+            return result
+        except (json.JSONDecodeError, ValueError, TypeError):
+            attempts += 1
+
+    return ClassificationResult(
+        needs_action=False,
+        category="uncategorized",
+        priority="low",
+        task_description=None,
+        reasoning="System failed to classify after multiple attempts.",
+        confidence=0.0,
+    )
+
+
+def _clean_email(email: EmailData) -> EmailData:
+    from app.helpers.clean_email_html import clean_email_html
+    from app.helpers.extract_latest_message import extract_latest_message
+    from app.helpers.remove_disclaimer import remove_disclaimer
+
+    return EmailData(
+        subject=email.subject,
+        body=remove_disclaimer(clean_email_html(extract_latest_message(email.body))),
+    )
+
+
+def _normalize_result(data: dict[str, Any]) -> ClassificationResult:
+    needs_action = bool(data.get("needs_action", False))
+    category = str(data.get("category", "uncategorized") or "uncategorized").lower()
+    if category not in VALID_CATEGORIES:
+        category = "uncategorized"
+    priority = str(data.get("priority", "low") or "low").lower()
+    if priority not in VALID_PRIORITIES:
+        priority = "low"
+    task_description = data.get("task_description")
+    if task_description is not None:
+        task_description = str(task_description).strip() or None
+    if needs_action and not task_description:
+        raise ValueError("task_description required when needs_action is true")
+    reasoning = str(data.get("reasoning", "") or "").strip() or "No reasoning provided."
+    confidence_raw = data.get("confidence", 0.0)
+    confidence = max(0.0, min(1.0, float(confidence_raw)))
+    return ClassificationResult(
+        needs_action=needs_action,
+        category=category,
+        priority=priority,
+        task_description=task_description,
+        reasoning=reasoning,
+        confidence=confidence,
+    )