feat(07-02): singleton OpenAIProvider + GenericOpenAIProvider + MAX_AI_CHARS removal

- openai_provider.py: singleton self._client=AsyncOpenAI(...) in __init__ (D-07),
  _truncate() 60/40 smart truncation (D-13), removed MAX_AI_CHARS constant,
  changed __init__ signature to include context_chars, removed _client() method
- generic_openai_provider.py: new class, subclasses OpenAIProvider, conditional
  response_format={"type":"json_object"} on supports_json_mode flag (D-01/D-02),
  imports parse_classification + parse_suggestions from ai.utils (D-02 contract)
- ollama_provider.py: added context_chars kwarg with default 8000, passes through
- lmstudio_provider.py: added context_chars kwarg with default 8000, passes through
- classifier.py: removed MAX_AI_CHARS constant and text[:MAX_AI_CHARS] slices;
  truncation now handled inside each provider via _truncate()
- requirements.txt: bumped anthropic floor to >=0.95.0 (D-03 output_config support)

backend/ai/openai_provider.py

@@ -1,18 +1,39 @@
+from __future__ import annotations
+
 from openai import AsyncOpenAI
 from ai.base import AIProvider, ClassificationResult
 from ai.utils import parse_classification, parse_suggestions
 
-MAX_AI_CHARS = 8_000
-
 
 class OpenAIProvider(AIProvider):
-    def __init__(self, api_key: str, model: str = "gpt-4o", base_url=None):  # type: ignore[type-arg]
-        self._api_key = api_key
+    def __init__(
+        self,
+        api_key: str,
+        model: str = "gpt-4o",
+        base_url: str | None = None,
+        context_chars: int = 8000,
+    ):
+        self._api_key = api_key or "not-needed"
         self._model = model
         self._base_url = base_url
+        self._context_chars = context_chars
+        # Singleton: created once in __init__, reused for all calls on this instance.
+        # Do NOT recreate per API call — AsyncOpenAI wraps an httpx.AsyncClient
+        # that maintains a connection pool; recreating per call destroys pool reuse
+        # and forces a new TLS handshake per request (D-07 / RESEARCH.md).
+        self._client = AsyncOpenAI(api_key=self._api_key, base_url=self._base_url)
 
-    def _client(self) -> AsyncOpenAI:
-        return AsyncOpenAI(api_key=self._api_key or "placeholder", base_url=self._base_url)
+    def _truncate(self, text: str) -> str:
+        """D-13 smart truncation: first 60% + last 40% of context window.
+
+        Captures both document introduction and conclusion, which carry the
+        most topic signal for long documents.
+        """
+        if len(text) <= self._context_chars:
+            return text
+        head_len = int(self._context_chars * 0.6)
+        tail_len = self._context_chars - head_len
+        return text[:head_len] + "\n[...truncated...]\n" + text[-tail_len:]
 
     async def classify(
         self,
@@ -23,9 +44,9 @@ class OpenAIProvider(AIProvider):
         topics_str = ", ".join(existing_topics) if existing_topics else "(none yet)"
         user_msg = (
             f"Existing topics: [{topics_str}]\n\n"
-            f"Document text:\n{document_text[:MAX_AI_CHARS]}"
+            f"Document text:\n{self._truncate(document_text)}"
         )
-        response = await self._client().chat.completions.create(
+        response = await self._client.chat.completions.create(
             model=self._model,
             max_tokens=1024,
             messages=[
@@ -44,9 +65,9 @@ class OpenAIProvider(AIProvider):
         user_msg = (
             "Suggest 3-5 topic names for this document. "
             "Return ONLY valid JSON: {\"suggested_topics\": [\"topic1\", \"topic2\"]}\n\n"
-            f"Document text:\n{document_text[:MAX_AI_CHARS]}"
+            f"Document text:\n{self._truncate(document_text)}"
         )
-        response = await self._client().chat.completions.create(
+        response = await self._client.chat.completions.create(
             model=self._model,
             max_tokens=256,
             messages=[
@@ -59,7 +80,7 @@ class OpenAIProvider(AIProvider):
 
     async def health_check(self) -> bool:
         try:
-            await self._client().chat.completions.create(
+            await self._client.chat.completions.create(
                 model=self._model,
                 max_tokens=5,
                 messages=[{"role": "user", "content": "ping"}],
@@ -67,5 +88,3 @@ class OpenAIProvider(AIProvider):
             return True
         except Exception:
             return False
-
-