feat(memory_tree): consolidate module + add agentic walk tool + tests

docs/whatsapp-data-flow.md

@@ -61,7 +61,7 @@ The scanner write-path RPCs are registered as **internal-only** in [`src/core/al
 The agent surfaces are exclusively read-only:
 
 - [`src/openhuman/tools/impl/whatsapp_data/`](../src/openhuman/tools/impl/whatsapp_data/) — `whatsapp_data_list_chats`, `whatsapp_data_list_messages`, `whatsapp_data_search_messages`. All three wrap their RPC counterparts and emit a `"provider": "whatsapp"` tag in the response so the agent can cite WhatsApp as the source.
-- [`src/openhuman/tools/impl/memory/tree/`](../src/openhuman/tools/impl/memory/tree/) — generic `memory_tree_*` tools. Filter by `source_kind: "chat"` or query directly; WhatsApp chat-day transcripts are tagged `whatsapp` so they surface in cross-source flows.
+- [`src/openhuman/memory_tree/tools/`](../src/openhuman/memory_tree/tools/) — generic `memory_tree_*` tools. Filter by `source_kind: "chat"` or query directly; WhatsApp chat-day transcripts are tagged `whatsapp` so they surface in cross-source flows.
 
 ## Why the orchestrator only lists three of these
 

src/bin/gmail_backfill_3d.rs

@@ -41,11 +41,11 @@ use openhuman_core::openhuman::composio::providers::registry::{
     get_provider, init_default_providers,
 };
 use openhuman_core::openhuman::config::Config;
-use openhuman_core::openhuman::memory::tree::content_store::read::{
+use openhuman_core::openhuman::memory_tree::content_store::read::{
     verify_chunk_file, verify_summary_file, VerifyResult,
 };
-use openhuman_core::openhuman::memory::tree::jobs::drain_until_idle;
-use openhuman_core::openhuman::memory::tree::store::{
+use openhuman_core::openhuman::memory_tree::jobs::drain_until_idle;
+use openhuman_core::openhuman::memory_tree::store::{
     get_chunk_content_pointers, list_chunks, list_summaries_with_content_path, ListChunksQuery,
 };
 

src/bin/memory_tree_init_smoke.rs

@@ -30,7 +30,7 @@ use std::sync::atomic::{AtomicUsize, Ordering};
 use std::sync::Arc;
 
 use openhuman_core::openhuman::config::Config;
-use openhuman_core::openhuman::memory::tree::store::with_connection;
+use openhuman_core::openhuman::memory_tree::store::with_connection;
 
 fn main() -> ExitCode {
     env_logger::Builder::from_env(env_logger::Env::default().default_filter_or("info"))

src/bin/slack_backfill.rs

@@ -211,8 +211,8 @@ async fn main() -> Result<()> {
 
     if cli.seal_probe {
         use chrono::{Duration, Utc};
-        use openhuman_core::openhuman::memory::tree::canonicalize::chat::{ChatBatch, ChatMessage};
-        use openhuman_core::openhuman::memory::tree::ingest::ingest_chat;
+        use openhuman_core::openhuman::memory_tree::canonicalize::chat::{ChatBatch, ChatMessage};
+        use openhuman_core::openhuman::memory_tree::ingest::ingest_chat;
 
         let connection_id = cli.connection_id.clone().ok_or_else(|| {
             anyhow::anyhow!(

src/core/all.rs

@@ -185,9 +185,9 @@ fn build_registered_controllers() -> Vec<RegisteredController> {
     // Document and knowledge graph storage
     controllers.extend(crate::openhuman::memory::all_memory_registered_controllers());
     // Memory tree ingestion layer (#707 — canonicalised chunks with provenance)
-    controllers.extend(crate::openhuman::memory::all_memory_tree_registered_controllers());
+    controllers.extend(crate::openhuman::memory_tree::all_memory_tree_registered_controllers());
     // Memory tree retrieval layer (#710 — LLM-callable read tools over the tree)
-    controllers.extend(crate::openhuman::memory::all_retrieval_registered_controllers());
+    controllers.extend(crate::openhuman::memory_tree::all_retrieval_registered_controllers());
     // Slack → memory-tree ingestion engine (per-message ingest, no bucketing)
     controllers.extend(
         crate::openhuman::composio::providers::slack::all_slack_memory_registered_controllers(),
@@ -226,8 +226,7 @@ fn build_registered_controllers() -> Vec<RegisteredController> {
     // Core binary update management
     controllers.extend(crate::openhuman::update::all_update_registered_controllers());
     // Hierarchical knowledge summarization
-    controllers
-        .extend(crate::openhuman::tree_summarizer::all_tree_summarizer_registered_controllers());
+    controllers.extend(crate::openhuman::memory_tree::all_tree_summarizer_registered_controllers());
     // Self-learning and user context enrichment
     controllers.extend(crate::openhuman::learning::all_learning_registered_controllers());
     // Conversation thread and message management
@@ -314,8 +313,8 @@ fn build_declared_controller_schemas() -> Vec<ControllerSchema> {
     schemas.extend(crate::openhuman::tools::all_tools_controller_schemas());
     schemas.extend(crate::openhuman::tool_registry::all_tool_registry_controller_schemas());
     schemas.extend(crate::openhuman::memory::all_memory_controller_schemas());
-    schemas.extend(crate::openhuman::memory::all_memory_tree_controller_schemas());
-    schemas.extend(crate::openhuman::memory::all_retrieval_controller_schemas());
+    schemas.extend(crate::openhuman::memory_tree::all_memory_tree_controller_schemas());
+    schemas.extend(crate::openhuman::memory_tree::all_retrieval_controller_schemas());
     schemas.extend(
         crate::openhuman::composio::providers::slack::all_slack_memory_controller_schemas(),
     );
@@ -333,7 +332,7 @@ fn build_declared_controller_schemas() -> Vec<ControllerSchema> {
     schemas.extend(crate::openhuman::subconscious::all_subconscious_controller_schemas());
     schemas.extend(crate::openhuman::webhooks::all_webhooks_controller_schemas());
     schemas.extend(crate::openhuman::update::all_update_controller_schemas());
-    schemas.extend(crate::openhuman::tree_summarizer::all_tree_summarizer_controller_schemas());
+    schemas.extend(crate::openhuman::memory_tree::all_tree_summarizer_controller_schemas());
     schemas.extend(crate::openhuman::learning::all_learning_controller_schemas());
     // Conversation thread and message management
     schemas.extend(crate::openhuman::threads::all_threads_controller_schemas());

src/core/cli.rs

@@ -68,7 +68,7 @@ pub fn run_from_cli_args(args: &[String]) -> Result<()> {
         }
         "text-input" => crate::openhuman::text_input::cli::run_text_input_command(&args[1..]),
         "tree-summarizer" => {
-            crate::openhuman::tree_summarizer::cli::run_tree_summarizer_command(&args[1..])
+            crate::openhuman::memory_tree::summarizer::cli::run_tree_summarizer_command(&args[1..])
         }
         "memory" => crate::core::memory_cli::run_memory_command(&args[1..]),
         "agent" => {

src/core/jsonrpc.rs

@@ -1558,7 +1558,7 @@ fn register_domain_subscribers(
             );
         }
 
-        crate::openhuman::memory::tree::jobs::start(config.clone());
+        crate::openhuman::memory_tree::jobs::start(config.clone());
 
         // Restart requests go through a subscriber so every trigger path shares
         // the same respawn logic.

src/openhuman/agent/harness/archivist.rs

@@ -24,17 +24,17 @@ use crate::openhuman::memory::store::profile::{self, FacetType};
 use crate::openhuman::memory::store::segments::{
     self, BoundaryConfig, BoundaryDecision, ConversationSegment,
 };
-use crate::openhuman::memory::tree::canonicalize::chat::{ChatBatch, ChatMessage};
-use crate::openhuman::memory::tree::chat::{ChatConsumer, ChatProvider};
-use crate::openhuman::memory::tree::ingest;
-use crate::openhuman::memory::tree::score::embed::{build_embedder_from_config, Embedder};
-use crate::openhuman::memory::tree::tree_source::summariser::llm::{
+use crate::openhuman::memory_tree::canonicalize::chat::{ChatBatch, ChatMessage};
+use crate::openhuman::memory_tree::chat::{ChatConsumer, ChatProvider};
+use crate::openhuman::memory_tree::ingest;
+use crate::openhuman::memory_tree::score::embed::{build_embedder_from_config, Embedder};
+use crate::openhuman::memory_tree::tree_source::summariser::llm::{
     LlmSummariser, LlmSummariserConfig,
 };
-use crate::openhuman::memory::tree::tree_source::summariser::{
+use crate::openhuman::memory_tree::tree_source::summariser::{
     Summariser, SummaryContext, SummaryInput,
 };
-use crate::openhuman::memory::tree::tree_source::types::TreeKind;
+use crate::openhuman::memory_tree::tree_source::types::TreeKind;
 use async_trait::async_trait;
 use parking_lot::Mutex;
 use rusqlite::Connection;
@@ -98,7 +98,7 @@ impl ArchivistHook {
     pub fn with_config(mut self, config: Config) -> Self {
         // Build the LLM chat provider for segment recap.
         let chat_provider: Option<Arc<dyn ChatProvider>> =
-            match crate::openhuman::memory::tree::chat::build_chat_provider(
+            match crate::openhuman::memory_tree::chat::build_chat_provider(
                 &config,
                 ChatConsumer::Summarise,
             ) {
@@ -650,7 +650,7 @@ impl ArchivistHook {
             .iter()
             .filter(|e| !e.content.trim().is_empty())
             .map(|e| {
-                use crate::openhuman::memory::tree::types::approx_token_count;
+                use crate::openhuman::memory_tree::types::approx_token_count;
                 let content = e.content.clone();
                 let token_count = approx_token_count(&content);
                 let ts = chrono::DateTime::from_timestamp(e.timestamp as i64, 0)

src/openhuman/agent/harness/archivist_tests.rs

@@ -1,7 +1,7 @@
 use super::*;
 use crate::openhuman::agent::hooks::{ToolCallRecord, TurnContext};
 use crate::openhuman::memory::store::{events as ev, fts5, segments as seg};
-use crate::openhuman::memory::tree::chat::ChatPrompt;
+use crate::openhuman::memory_tree::chat::ChatPrompt;
 
 fn setup_conn() -> Arc<Mutex<Connection>> {
     let conn = Connection::open_in_memory().unwrap();
@@ -343,7 +343,7 @@ async fn phase0_episodic_rows_and_segment_without_learning_enabled() {
 struct StubChatProvider;
 
 #[async_trait::async_trait]
-impl crate::openhuman::memory::tree::chat::ChatProvider for StubChatProvider {
+impl crate::openhuman::memory_tree::chat::ChatProvider for StubChatProvider {
     fn name(&self) -> &str {
         "stub:test"
     }
@@ -361,7 +361,7 @@ impl crate::openhuman::memory::tree::chat::ChatProvider for StubChatProvider {
 struct StubEmbedder;
 
 #[async_trait::async_trait]
-impl crate::openhuman::memory::tree::score::embed::Embedder for StubEmbedder {
+impl crate::openhuman::memory_tree::score::embed::Embedder for StubEmbedder {
     fn name(&self) -> &'static str {
         "stub-embedder-v1"
     }
@@ -546,7 +546,7 @@ async fn phase1_flush_open_segment_finalizes_trailing_segment() {
 //   g) flush_open_segment also triggers tree ingest.
 
 use crate::openhuman::config::Config;
-use crate::openhuman::memory::tree::store::{count_chunks, list_chunks, ListChunksQuery};
+use crate::openhuman::memory_tree::store::{count_chunks, list_chunks, ListChunksQuery};
 use tempfile::TempDir;
 
 /// Build a Config that points at a temp workspace, suitable for tree-ingest tests.

src/openhuman/agent/harness/payload_summarizer.rs

@@ -309,7 +309,7 @@ impl PayloadSummarizer for SubagentPayloadSummarizer {
 }
 
 /// Rough token estimate: ~4 characters per token. Mirrors
-/// [`crate::openhuman::tree_summarizer::types::estimate_tokens`] but
+/// [`crate::openhuman::memory_tree::summarizer::types::estimate_tokens`] but
 /// returns `usize` (not `u32`) and lives here to avoid a cross-module
 /// dependency from the agent harness on the tree summarizer.
 fn estimate_tokens(text: &str) -> usize {

src/openhuman/agent/harness/session/turn.rs

@@ -2150,7 +2150,7 @@ impl Agent {
 }
 
 /// Wrapper around
-/// [`crate::openhuman::tree_summarizer::store::collect_root_summaries_with_caps`]
+/// [`crate::openhuman::memory_tree::summarizer::store::collect_root_summaries_with_caps`]
 /// that takes user-resolved per-namespace and total caps. The actual
 /// limits are derived from the active
 /// [`crate::openhuman::config::schema::agent::MemoryContextWindow`]
@@ -2160,7 +2160,7 @@ fn collect_tree_root_summaries(
     per_namespace_cap: usize,
     total_cap: usize,
 ) -> Vec<(String, String)> {
-    crate::openhuman::tree_summarizer::store::collect_root_summaries_with_caps(
+    crate::openhuman::memory_tree::summarizer::store::collect_root_summaries_with_caps(
         workspace_dir,
         per_namespace_cap,
         total_cap,

src/openhuman/agent/harness/subagent_runner/handoff.rs

@@ -29,7 +29,7 @@ use std::sync::Mutex as StdMutex;
 /// cache instead of being pushed into history raw. Token count is
 /// estimated at ~4 chars/token (mirrors
 /// `crate::openhuman::agent::harness::payload_summarizer` and
-/// `crate::openhuman::tree_summarizer::types::estimate_tokens`).
+/// `crate::openhuman::memory_tree::summarizer::types::estimate_tokens`).
 ///
 /// Set at `50_000` so the clean Gmail / Notion envelopes emitted by provider
 /// post-processing fit through unchanged for normal workloads — only

src/openhuman/agent/tree_loader.rs

@@ -21,7 +21,7 @@
 //! concatenate without branching.
 
 use crate::openhuman::config::Config;
-use crate::openhuman::memory::tree::retrieval::query_global;
+use crate::openhuman::memory_tree::retrieval::query_global;
 
 /// Default lookback window for the eager digest. Mirrors the language in
 /// the orchestrator prompt ("7-day digest pre-loaded into session context").

src/openhuman/channels/runtime/startup.rs

@@ -587,7 +587,7 @@ pub async fn start_channels(config: Config) -> Result<()> {
     };
     // Register the tree summarizer event subscriber for observability logging.
     let _tree_summarizer_handle = bus.subscribe(Arc::new(
-        crate::openhuman::tree_summarizer::bus::TreeSummarizerEventSubscriber::new(),
+        crate::openhuman::memory_tree::summarizer::bus::TreeSummarizerEventSubscriber::new(),
     ));
 
     let max_in_flight_messages = compute_max_in_flight_messages(channels.len());

src/openhuman/composio/ops_test.rs

@@ -576,8 +576,8 @@ async fn composio_execute_via_mock_propagates_backend_error() {
 #[tokio::test]
 async fn composio_sync_gmail_via_mock_archives_raw_email_and_updates_outcome() {
     use crate::openhuman::config::TEST_ENV_LOCK;
-    use crate::openhuman::memory::tree::content_store::raw::{raw_rel_path, RawKind};
-    use crate::openhuman::memory::tree::rpc::{list_chunks_rpc, ListChunksRequest};
+    use crate::openhuman::memory_tree::content_store::raw::{raw_rel_path, RawKind};
+    use crate::openhuman::memory_tree::rpc::{list_chunks_rpc, ListChunksRequest};
     let _cache_guard = CACHE_TEST_GUARD.lock().unwrap_or_else(|e| e.into_inner());
     let _env_guard = TEST_ENV_LOCK.lock().unwrap_or_else(|e| e.into_inner());
 

src/openhuman/composio/providers/gmail/ingest.rs

@@ -21,16 +21,14 @@ use anyhow::Result;
 use serde_json::Value;
 
 use crate::openhuman::config::Config;
-use crate::openhuman::memory::tree::canonicalize::email::{EmailMessage, EmailThread};
-use crate::openhuman::memory::tree::canonicalize::email_clean::{
-    extract_email, parse_message_date,
-};
-use crate::openhuman::memory::tree::content_store::raw::{
+use crate::openhuman::memory_tree::canonicalize::email::{EmailMessage, EmailThread};
+use crate::openhuman::memory_tree::canonicalize::email_clean::{extract_email, parse_message_date};
+use crate::openhuman::memory_tree::content_store::raw::{
     self as raw_store, raw_rel_path, slug_account_email, RawItem, RawKind,
 };
-use crate::openhuman::memory::tree::ingest::{ingest_email, IngestResult};
-use crate::openhuman::memory::tree::store::{set_chunk_raw_refs, RawRef};
-use crate::openhuman::memory::tree::util::redact::redact;
+use crate::openhuman::memory_tree::ingest::{ingest_email, IngestResult};
+use crate::openhuman::memory_tree::store::{set_chunk_raw_refs, RawRef};
+use crate::openhuman::memory_tree::util::redact::redact;
 
 /// Provider name embedded in the canonical email-thread header. Matches
 /// the value `memory::tree::retrieval::source::PLATFORM_KINDS` expects.

src/openhuman/composio/providers/profile.rs

@@ -32,7 +32,7 @@ use std::collections::BTreeMap;
 
 /// Shape of an identifier persisted against a connection. Mirrors the
 /// matching dimensions of the memory tree's
-/// `crate::openhuman::memory::tree::score::extract::EntityKind` so the
+/// `crate::openhuman::memory_tree::score::extract::EntityKind` so the
 /// self-check is a direct `(toolkit, kind, value)` lookup.
 #[derive(Debug, Clone, Copy, PartialEq, Eq)]
 pub enum IdentityKind {

src/openhuman/composio/providers/slack/ingest.rs

@@ -2,7 +2,7 @@
 //!
 //! Owns the conversion from a page of [`SlackMessage`]s (post-processed
 //! and enriched by [`super::sync`]) into per-channel [`ChatBatch`]es and
-//! drives [`memory::tree::ingest::ingest_chat`] per message.
+//! drives [`memory_tree::ingest::ingest_chat`] per message.
 //!
 //! ## Source-id scope
 //!
@@ -30,16 +30,16 @@ use anyhow::Result;
 
 use super::types::SlackMessage;
 use crate::openhuman::config::Config;
-use crate::openhuman::memory::tree::canonicalize::chat::{ChatBatch, ChatMessage};
-use crate::openhuman::memory::tree::content_store::raw::{
+use crate::openhuman::memory_tree::canonicalize::chat::{ChatBatch, ChatMessage};
+use crate::openhuman::memory_tree::content_store::raw::{
     self as raw_store, raw_rel_path, RawItem, RawKind,
 };
-use crate::openhuman::memory::tree::ingest::ingest_chat;
-use crate::openhuman::memory::tree::store::{set_chunk_raw_refs, RawRef};
-use crate::openhuman::memory::tree::util::redact::redact;
+use crate::openhuman::memory_tree::ingest::ingest_chat;
+use crate::openhuman::memory_tree::store::{set_chunk_raw_refs, RawRef};
+use crate::openhuman::memory_tree::util::redact::redact;
 
 /// Platform identifier embedded in the canonical chat transcript header.
-/// Matches the value `memory::tree::retrieval::source::PLATFORM_KINDS` expects.
+/// Matches the value `memory_tree::retrieval::source::PLATFORM_KINDS` expects.
 pub const SLACK_PLATFORM: &str = "slack";
 
 /// Tags attached to every Slack-ingested chunk. Stable list — retrieval

src/openhuman/config/ops.rs

@@ -573,7 +573,7 @@ pub async fn apply_model_settings(
     // so a UI embedder switch recovers prior memory under the new
     // signature. Coverage-gated + non-fatal: if the active signature did
     // not actually change, this enqueues nothing.
-    crate::openhuman::memory::tree::jobs::ensure_reembed_backfill(config);
+    crate::openhuman::memory_tree::jobs::ensure_reembed_backfill(config);
     let snapshot = snapshot_config_json(config)?;
     Ok(RpcOutcome::new(
         snapshot,
@@ -623,7 +623,7 @@ pub async fn apply_memory_settings(
     // dark. Idempotent + non-fatal (covered space enqueues nothing; errors
     // are logged, never fail the settings save). §7's migration is
     // one-shot so it does not cover a later switch — this does.
-    crate::openhuman::memory::tree::jobs::ensure_reembed_backfill(config);
+    crate::openhuman::memory_tree::jobs::ensure_reembed_backfill(config);
     let snapshot = snapshot_config_json(config)?;
     Ok(RpcOutcome::new(
         snapshot,

src/openhuman/context/segment_recap_summarizer_tests.rs

@@ -16,7 +16,7 @@ use crate::openhuman::agent::hooks::{PostTurnHook as _, TurnContext};
 use crate::openhuman::context::summarizer::{Summarizer, SummaryStats};
 use crate::openhuman::inference::provider::{ChatMessage, ConversationMessage};
 use crate::openhuman::memory::store::{fts5, segments as seg};
-use crate::openhuman::memory::tree::chat::ChatPrompt;
+use crate::openhuman::memory_tree::chat::ChatPrompt;
 use anyhow::Result;
 use async_trait::async_trait;
 use parking_lot::Mutex;
@@ -40,7 +40,7 @@ fn setup_conn() -> Arc<Mutex<Connection>> {
 struct StubChatProvider;
 
 #[async_trait]
-impl crate::openhuman::memory::tree::chat::ChatProvider for StubChatProvider {
+impl crate::openhuman::memory_tree::chat::ChatProvider for StubChatProvider {
     fn name(&self) -> &str {
         "stub:test"
     }
@@ -56,7 +56,7 @@ impl crate::openhuman::memory::tree::chat::ChatProvider for StubChatProvider {
 struct FailingChatProvider;
 
 #[async_trait]
-impl crate::openhuman::memory::tree::chat::ChatProvider for FailingChatProvider {
+impl crate::openhuman::memory_tree::chat::ChatProvider for FailingChatProvider {
     fn name(&self) -> &str {
         "stub:failing"
     }
@@ -72,7 +72,7 @@ impl crate::openhuman::memory::tree::chat::ChatProvider for FailingChatProvider
 struct StubEmbedder;
 
 #[async_trait]
-impl crate::openhuman::memory::tree::score::embed::Embedder for StubEmbedder {
+impl crate::openhuman::memory_tree::score::embed::Embedder for StubEmbedder {
     fn name(&self) -> &'static str {
         "stub-embedder-v1"
     }

src/openhuman/doctor/core.rs

@@ -817,7 +817,7 @@ fn check_memory_tree_db(config: &Config, items: &mut Vec<DiagnosticItem>) {
     }
 
     // ── Probe connection ─────────────────────────────────────────────
-    match crate::openhuman::memory::tree::store::with_connection(config, |conn| {
+    match crate::openhuman::memory_tree::store::with_connection(config, |conn| {
         let n: i64 = conn.query_row("SELECT COUNT(*) FROM mem_tree_chunks", [], |r| r.get(0))?;
         Ok(n)
     }) {