🐾 feat(library): add 'email' library type

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>

.env.example

@@ -91,6 +91,19 @@ SPELUNKER_S3_REGION_NAME=us-east-1
 SPELUNKER_S3_USE_SSL=True
 SPELUNKER_S3_VERIFY=True
 
+# --- Kairos S3 (cross-bucket reads for ingest, source="kairos-mail") ---
+# Consumed by: worker only
+# Kairos renders synced mail to text documents in its own bucket and posts
+# ingest requests with source="kairos-mail". These creds should be scoped
+# read-only to the Kairos bucket in your secret manager.
+KAIROS_S3_ENDPOINT_URL=https://nyx.helu.ca:8555
+KAIROS_S3_ACCESS_KEY_ID=
+KAIROS_S3_SECRET_ACCESS_KEY=
+KAIROS_S3_BUCKET_NAME=kairos
+KAIROS_S3_REGION_NAME=us-east-1
+KAIROS_S3_USE_SSL=True
+KAIROS_S3_VERIFY=True
+
 # --- Celery / RabbitMQ (Oberon) ---------------------------------------------
 # Consumed by: app (producer), worker (consumer). NOT mcp.
 # Remember to percent-encode any password characters that have meaning in a

docker-compose.yaml

@@ -346,6 +346,13 @@ services:
       - SPELUNKER_S3_REGION_NAME=${SPELUNKER_S3_REGION_NAME}
       - SPELUNKER_S3_USE_SSL=${SPELUNKER_S3_USE_SSL}
       - SPELUNKER_S3_VERIFY=${SPELUNKER_S3_VERIFY}
+      - KAIROS_S3_ENDPOINT_URL=${KAIROS_S3_ENDPOINT_URL}
+      - KAIROS_S3_ACCESS_KEY_ID=${KAIROS_S3_ACCESS_KEY_ID}
+      - KAIROS_S3_SECRET_ACCESS_KEY=${KAIROS_S3_SECRET_ACCESS_KEY}
+      - KAIROS_S3_BUCKET_NAME=${KAIROS_S3_BUCKET_NAME}
+      - KAIROS_S3_REGION_NAME=${KAIROS_S3_REGION_NAME}
+      - KAIROS_S3_USE_SSL=${KAIROS_S3_USE_SSL}
+      - KAIROS_S3_VERIFY=${KAIROS_S3_VERIFY}
       # Celery / RabbitMQ
       - CELERY_BROKER_URL=${CELERY_BROKER_URL}
       - CELERY_RESULT_BACKEND=${CELERY_RESULT_BACKEND}

mnemosyne/library/api/serializers.py

@@ -15,6 +15,7 @@ LIBRARY_TYPE_CHOICES = [
     "film",
     "art",
     "journal",
+    "email",
     "business",
     "finance",
 ]

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,

mnemosyne/library/models.py

@@ -77,6 +77,7 @@ class Library(StructuredNode):
             "film": "Film",
             "art": "Art",
             "journal": "Journal",
+            "email": "Email",
             "business": "Business",
             "finance": "Finance",
         },

mnemosyne/library/tests/test_content_types.py

@@ -21,6 +21,7 @@ class LibraryTypeDefaultsTests(TestCase):
         "film",
         "art",
         "journal",
+        "email",
         "business",
         "finance",
     }

mnemosyne/mnemosyne/settings.py

@@ -268,6 +268,15 @@ SPELUNKER_S3_REGION_NAME = env("SPELUNKER_S3_REGION_NAME", default="us-east-1")
 SPELUNKER_S3_USE_SSL = env.bool("SPELUNKER_S3_USE_SSL", default=False)
 SPELUNKER_S3_VERIFY = env.bool("SPELUNKER_S3_VERIFY", default=True)
 
+# Kairos renders synced mail to text documents in its own bucket.
+KAIROS_S3_ENDPOINT_URL = env("KAIROS_S3_ENDPOINT_URL", default="")
+KAIROS_S3_ACCESS_KEY_ID = env("KAIROS_S3_ACCESS_KEY_ID", default="")
+KAIROS_S3_SECRET_ACCESS_KEY = env("KAIROS_S3_SECRET_ACCESS_KEY", default="")
+KAIROS_S3_BUCKET_NAME = env("KAIROS_S3_BUCKET_NAME", default="kairos")
+KAIROS_S3_REGION_NAME = env("KAIROS_S3_REGION_NAME", default="us-east-1")
+KAIROS_S3_USE_SSL = env.bool("KAIROS_S3_USE_SSL", default=False)
+KAIROS_S3_VERIFY = env.bool("KAIROS_S3_VERIFY", default=True)
+
 # Registry keyed by the ingest `source` field. Unknown/blank sources fall
 # back to "daedalus" for backwards compatibility.
 SOURCE_S3_BUCKETS = {
@@ -289,6 +298,15 @@ SOURCE_S3_BUCKETS = {
         "use_ssl": SPELUNKER_S3_USE_SSL,
         "verify": SPELUNKER_S3_VERIFY,
     },
+    "kairos-mail": {
+        "endpoint_url": KAIROS_S3_ENDPOINT_URL,
+        "access_key_id": KAIROS_S3_ACCESS_KEY_ID,
+        "secret_access_key": KAIROS_S3_SECRET_ACCESS_KEY,
+        "bucket_name": KAIROS_S3_BUCKET_NAME,
+        "region_name": KAIROS_S3_REGION_NAME,
+        "use_ssl": KAIROS_S3_USE_SSL,
+        "verify": KAIROS_S3_VERIFY,
+    },
 }
 
 # --- Celery / RabbitMQ ---