Data Models Reference

Overview

Beever Atlas uses a hierarchical knowledge model with three tiers:

Tier 0: Channel Summary - Overall channel narrative
Tier 1: Topic Clusters - Grouped related facts
Tier 2: Atomic Facts - Individual extracted knowledge units
Graph: Entities and Relationships - Knowledge graph

AtomicFact

A single extracted fact stored in Weaviate (Tier 2).

Fields

Field	Type	Description
`id`	string	Unique fact identifier (UUID v5)
`memory_text`	string	The extracted fact text
`quality_score`	float	Quality score 0.0-1.0
`tier`	string	Always `"atomic"` for facts
`cluster_id`	string \| null	Parent topic cluster ID
`channel_id`	string	Source channel identifier
`platform`	string	Platform name (e.g., "slack")
`author_id`	string	Author's platform user ID
`author_name`	string	Author's display name
`message_ts`	string	Message timestamp
`thread_ts`	string \| null	Thread root timestamp
`source_message_id`	string	Platform message ID
`topic_tags`	string[]	Topic category tags
`entity_tags`	string[]	Mentioned entity names
`action_tags`	string[]	Action item keywords
`importance`	string	`"low" \| "medium" \| "high"`
`graph_entity_ids`	string[]	Linked graph entity IDs
`source_media_urls`	string[]	URLs of attached media
`source_media_names`	string[]	Filenames of attached media
`source_link_urls`	string[]	URLs of linked resources
`source_link_titles`	string[]	Titles of linked resources
`source_link_descriptions`	string[]	Descriptions of linked resources
`valid_at`	datetime \| null	When fact became valid
`invalid_at`	datetime \| null	When fact became invalid
`superseded_by`	string \| null	ID of fact that replaced this
`supersedes`	string \| null	ID of fact this replaced
`potential_contradiction`	boolean	Flagged for conflict review
`text_vector`	float[] \| null	Embedding vector
`fact_type`	string	`"decision" \| "opinion" \| "observation" \| "action_item" \| "question"`
`thread_context_summary`	string	Thread deliberation summary
`source_lang`	string	BCP-47 language tag (e.g., "en", "zh-HK")

Example

{
  "id": "550e8400-e29b-41d4-a716-446655440000",
  "memory_text": "The team decided to use JWT tokens for authentication with a 1-hour expiration.",
  "quality_score": 0.85,
  "tier": "atomic",
  "cluster_id": "cluster_123",
  "channel_id": "C12345",
  "platform": "slack",
  "author_id": "U12345",
  "author_name": "Jane Smith",
  "message_ts": "2026-04-13T10:30:00Z",
  "thread_ts": null,
  "source_message_id": "1234567890.123456",
  "topic_tags": ["authentication", "security"],
  "entity_tags": ["JWT", "OAuth"],
  "action_tags": [],
  "importance": "high",
  "graph_entity_ids": ["entity_jwt_decision"],
  "fact_type": "decision",
  "source_lang": "en"
}

TopicCluster

A Tier 1 topic cluster grouping related atomic facts.

Fields

Field	Type	Description
`id`	string	Unique cluster identifier
`tier`	string	Always `"topic"`
`channel_id`	string	Channel identifier
`title`	string	Short descriptive name (5-10 words)
`summary`	string	Narrative of what happened (2-3 sentences)
`current_state`	string	Where things stand now (1-2 sentences)
`open_questions`	string	Unresolved tensions/debates
`impact_note`	string	Scope and significance
`topic_tags`	string[]	Topic category tags
`member_ids`	string[]	Member fact IDs
`member_count`	int	Number of member facts
`centroid_vector`	float[] \| null	Cluster center embedding
`created_at`	datetime	Creation timestamp
`updated_at`	datetime	Last update timestamp
`key_entities`	object[]	Related entities `[{id, name, type}]`
`key_relationships`	object[]	Key relationships `[{source, type, target, confidence}]`
`date_range_start`	string	Earliest fact timestamp
`date_range_end`	string	Latest fact timestamp
`authors`	string[]	Contributors to this topic
`media_refs`	string[]	Referenced media URLs
`media_names`	string[]	Referenced media filenames
`link_refs`	string[]	Referenced link URLs
`high_importance_count`	int	Number of high-importance facts
`related_cluster_ids`	string[]	Related topic clusters
`staleness_score`	float	0.0=fresh, 1.0=very stale
`status`	string	`"active" \| "completed" \| "stale"`
`fact_type_counts`	object	Counts by fact type `{decision: N, ...}`
`key_facts`	object[]	Top facts with metadata
`decisions`	object[]	Extracted decisions
`people`	object[]	Involved people `[{name, role, entity_id}]`
`technologies`	object[]	Technologies mentioned `[{name, category, champion}]`
`projects`	object[]	Projects referenced `[{name, status, owner, blockers}]`
`faq_candidates`	object[]	Q&A pairs `[{question, answer}]`

Example

{
  "id": "topic_123",
  "tier": "topic",
  "channel_id": "C12345",
  "title": "Authentication System Migration to JWT",
  "summary": "The team migrated from session-based auth to JWT tokens. This decision was driven by scalability needs and better mobile support.",
  "current_state": "Migration complete, JWT tokens in production with 1-hour expiration.",
  "open_questions": "Token refresh strategy still under review.",
  "impact_note": "Affects all API clients and requires re-authentication flow updates.",
  "status": "active",
  "member_count": 15,
  "staleness_score": 0.1
}

ChannelSummary

A Tier 0 channel-level summary consolidating all topic clusters.

Fields

Field	Type	Description
`id`	string	Unique summary identifier
`tier`	string	Always `"summary"`
`channel_id`	string	Channel identifier
`channel_name`	string	Resolved display name
`text`	string	Overall narrative overview
`description`	string	One-line channel purpose (max 200 chars)
`themes`	string	Main knowledge areas and interrelations
`momentum`	string	Active vs completed vs stale
`team_dynamics`	string	Decision drivers, collaboration patterns
`cluster_count`	int	Number of topic clusters
`fact_count`	int	Total number of facts
`updated_at`	datetime	Last update timestamp
`key_decisions`	object[]	Top decisions
`key_entities`	object[]	Top entities
`key_topics`	object[]	Top topics with metadata
`date_range_start`	string	Earliest content
`date_range_end`	string	Latest content
`media_count`	int	Total media references
`author_count`	int	Unique contributors
`worst_staleness`	float	Highest staleness in clusters
`top_decisions`	object[]	Key decisions across all topics
`top_people`	object[]	Key people by expertise
`tech_stack`	object[]	Technologies referenced
`active_projects`	object[]	Ongoing projects
`glossary_terms`	object[]	Domain terminology
`recent_activity_summary`	object	Activity metrics
`topic_graph_edges`	object[]	Topic relationships

GraphEntity

An entity node in the Neo4j knowledge graph.

Fields

Field	Type	Description
`id`	string	Unique entity identifier
`name`	string	Entity name
`type`	string	`"Person" \| "Decision" \| "Project" \| "Technology" \| "Team" \| "Meeting"`
`scope`	string	`"global" \| "channel"`
`channel_id`	string \| null	Channel scope (if channel-scoped)
`properties`	object	Additional entity properties
`aliases`	string[]	Alternative names
`status`	string	`"active" \| "pending"`
`pending_since`	datetime \| null	When status became pending
`name_vector`	float[] \| null	Name embedding for fuzzy matching
`source_fact_ids`	string[]	Source facts
`source_message_id`	string	Originating message
`message_ts`	string	Message timestamp
`source_lang`	string	BCP-47 language tag
`created_at`	datetime	Creation timestamp
`updated_at`	datetime	Last update timestamp

Entity Types

Person: Team members, stakeholders
Decision: Decisions made, conclusions reached
Project: Projects, initiatives, work streams
Technology: Tools, frameworks, libraries
Team: Teams, groups, departments
Meeting: Meetings, standups, reviews

GraphRelationship

A relationship edge in the Neo4j knowledge graph.

Fields

Field	Type	Description
`id`	string	Unique relationship identifier
`type`	string	Relationship type (see below)
`source`	string	Source entity name
`target`	string	Target entity name
`confidence`	float	Confidence score 0.0-1.0
`valid_from`	string \| null	When relationship became valid
`valid_until`	string \| null	When relationship became invalid
`context`	string	Relationship context
`source_message_id`	string	Originating message
`source_fact_id`	string	Source fact
`created_at`	datetime	Creation timestamp

Relationship Types

DECIDED_BY: Person decided Decision
PARTICIPATED_IN: Person participated in Meeting
WORKS_ON: Person works on Project
USES: Project uses Technology
MEMBER_OF: Person member of Team
RELATED_TO: General relationship

WikiPage

A generated wiki page with enhanced Markdown content.

Fields

Field	Type	Description
`id`	string	Page ID (`"overview"`, `"people"`, `"topic-xyz"`)
`slug`	string	URL slug
`title`	string	Page title
`page_type`	string	`"fixed" \| "topic" \| "sub-topic"`
`parent_id`	string \| null	Parent page ID
`section_number`	string	Section number (`"1"`, `"2.1"`, `"2.1.1"`)
`content`	string	Enhanced Markdown content
`summary`	string	1-2 sentence summary
`memory_count`	int	Number of source facts
`last_updated`	datetime	Last update timestamp
`citations`	WikiCitation[]	Source citations
`children`	WikiPageRef[]	Child pages

WikiPage Types

fixed: Standard pages (overview, people, decisions, etc.)
topic: Pages generated from topic clusters
sub-topic: Nested topic pages

WikiCitation

A source citation for wiki page content.

Fields

Field	Type	Description
`id`	string	Citation reference (`"[1]"`, `"[2]"`)
`author`	string	Message author
`channel`	string	Channel name
`timestamp`	string	Message timestamp
`text_excerpt`	string	First 100 characters
`permalink`	string	URL to original message
`media_type`	string \| null	`"pdf" \| "image" \| "link" \| "video" \| "audio"`
`media_name`	string \| null	Filename or domain

EntityKnowledgeCard

Cross-channel aggregation of all knowledge about a single graph entity.

Fields

Field	Type	Description
`id`	string	Unique card identifier
`tier`	string	Always `"entity_card"`
`entity_id`	string	Graph entity ID
`entity_name`	string	Entity name
`entity_type`	string	Entity type
`channel_ids`	string[]	Channels where entity appears
`cluster_ids`	string[]	Related topic clusters
`fact_count`	int	Number of related facts
`fact_type_breakdown`	object	Fact counts by type
`key_facts`	string[]	Most important facts
`related_entities`	object[]	Related entities
`last_mentioned_at`	string	Latest mention timestamp
`staleness_score`	float	0.0=fresh, 1.0=stale
`summary`	string	Entity summary
`updated_at`	datetime	Last update timestamp

Model Relationships

ChannelSummary (Tier 0)
    ├─> TopicCluster[] (Tier 1)
    │       └─> AtomicFact[] (Tier 2)
    │               └─> GraphEntity (via graph_entity_ids)
    │                       └─> GraphRelationship
    └─> WikiPage
            └─> WikiCitation (cites AtomicFact)

Storage

Weaviate: AtomicFact, TopicCluster (Tier 1-2)
MongoDB: WikiPage, WikiMetadata, ChannelSummary
Neo4j: GraphEntity, GraphRelationship (knowledge graph)

On this page