From 9f20110f56f93c94ab10a0ca2c3a54bfea82fa68 Mon Sep 17 00:00:00 2001
From: Robert Helewka <r@helu.ca>
Date: Sun, 12 Jul 2026 08:03:28 -0400
Subject: [PATCH] =?UTF-8?q?=F0=9F=90=BE=20feat(library):=20add=20'email'?=
 =?UTF-8?q?=20library=20type?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Kairos is about to provision per-user app-managed (workspace) libraries for
synced mail, and none of the existing types fit correspondence. New entry in
LIBRARY_TYPE_DEFAULTS mirrors journal's entry-level chunking (one message ≈
one entry) with instructions tuned for email: correspondents, dates,
requests/commitments, and quoted-reply context. Registered in the Library
node choices and the API serializer choice list; forms and
load_library_types derive from the registry, and the compose init sidecar
re-runs the seeder on deploy.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
---
 mnemosyne/library/api/serializers.py          |  1 +
 mnemosyne/library/content_types.py            | 34 ++++++++++++++++++-
 mnemosyne/library/models.py                   |  1 +
 mnemosyne/library/tests/test_content_types.py |  1 +
 4 files changed, 36 insertions(+), 1 deletion(-)

diff --git a/mnemosyne/library/api/serializers.py b/mnemosyne/library/api/serializers.py
index c4cb03c..b28ce9c 100644
--- a/mnemosyne/library/api/serializers.py
+++ b/mnemosyne/library/api/serializers.py
@@ -15,6 +15,7 @@ LIBRARY_TYPE_CHOICES = [
     "film",
     "art",
     "journal",
+    "email",
     "business",
     "finance",
 ]
diff --git a/mnemosyne/library/content_types.py b/mnemosyne/library/content_types.py
index 215df02..40369d4 100644
--- a/mnemosyne/library/content_types.py
+++ b/mnemosyne/library/content_types.py
@@ -241,6 +241,38 @@ LIBRARY_TYPE_DEFAULTS = {
             "4) The commercial purpose — positioning, pricing, capability demonstration."
         ),
     },
+    "email": {
+        "chunking_config": {
+            "strategy": "entry_level",
+            "chunk_size": 512,
+            "chunk_overlap": 32,
+            "respect_boundaries": ["message", "quote", "paragraph"],
+        },
+        "embedding_instruction": (
+            "Represent this email message for retrieval. "
+            "Focus on the sender, recipients, subject, dates, requests and "
+            "commitments made, and the people, organizations, and events discussed."
+        ),
+        "reranker_instruction": (
+            "Re-rank email messages based on relevance to the query. "
+            "Prioritize messages matching the correspondents, subject matter, "
+            "time period, and any specific commitments or requests mentioned."
+        ),
+        "llm_context_prompt": (
+            "The following excerpts are from personal email correspondence. "
+            "This is private content — answer with discretion. Attribute "
+            "statements to their senders, note dates, and distinguish what was "
+            "asked from what was agreed. Quoted text below a reply is earlier "
+            "context, not the sender's own words."
+        ),
+        "vision_prompt": (
+            "Analyze this image from an email message. Identify:\n"
+            "1) Image type (photograph, screenshot, scanned document, chart, signature graphic).\n"
+            "2) What it depicts — people, places, documents, data.\n"
+            "3) Any visible text, dates, or figures.\n"
+            "4) Its role in the message — attachment content, inline illustration, or boilerplate."
+        ),
+    },
     "finance": {
         "chunking_config": {
             "strategy": "section_aware",
@@ -282,7 +314,7 @@ def get_library_type_config(library_type):
 
     Args:
         library_type: One of 'fiction', 'nonfiction', 'technical', 'music',
-                      'film', 'art', 'journal', 'business', 'finance'
+                      'film', 'art', 'journal', 'email', 'business', 'finance'
 
     Returns:
         dict with keys: chunking_config, embedding_instruction,
diff --git a/mnemosyne/library/models.py b/mnemosyne/library/models.py
index 69fe482..2594d35 100644
--- a/mnemosyne/library/models.py
+++ b/mnemosyne/library/models.py
@@ -77,6 +77,7 @@ class Library(StructuredNode):
             "film": "Film",
             "art": "Art",
             "journal": "Journal",
+            "email": "Email",
             "business": "Business",
             "finance": "Finance",
         },
diff --git a/mnemosyne/library/tests/test_content_types.py b/mnemosyne/library/tests/test_content_types.py
index 8725c55..4e5c3f1 100644
--- a/mnemosyne/library/tests/test_content_types.py
+++ b/mnemosyne/library/tests/test_content_types.py
@@ -21,6 +21,7 @@ class LibraryTypeDefaultsTests(TestCase):
         "film",
         "art",
         "journal",
+        "email",
         "business",
         "finance",
     }