sqlite_graphrag/commands/

remember.rs

//! Handler for the `remember` CLI subcommand.

use crate::chunking;
use crate::cli::MemoryType;
use crate::entity_type::EntityType;
use crate::errors::AppError;
use crate::i18n::errors_msg;
use crate::output::{self, JsonOutputFormat, RememberResponse};
use crate::paths::AppPaths;
use crate::storage::chunks as storage_chunks;
use crate::storage::connection::{ensure_schema, open_rw};
use crate::storage::entities::{NewEntity, NewRelationship};
use crate::storage::memories::NewMemory;
use crate::storage::{entities, memories, urls as storage_urls, versions};
use serde::Deserialize;

/// Returns the number of rows that will be written to `memory_chunks` for the
/// given chunk count. Single-chunk bodies are stored directly in the
/// `memories` row, so no chunk row is appended (returns `0`). Multi-chunk
/// bodies persist every chunk and the count equals `chunks_created`.
///
/// Centralized as a function so the H-M8 invariant is unit-testable without
/// running the full handler. The schema for `chunks_persisted` documents this
/// contract explicitly (see `docs/schemas/remember.schema.json`).
fn compute_chunks_persisted(chunks_created: usize) -> usize {
    if chunks_created > 1 {
        chunks_created
    } else {
        0
    }
}

#[derive(clap::Args)]
#[command(after_long_help = "EXAMPLES:\n  \
    # Create a memory with inline body\n  \
    sqlite-graphrag remember --name design-auth --type decision \\\n    \
    --description \"auth design\" --body \"JWT for stateless auth\"\n\n  \
    # Create with curated graph via --graph-stdin\n  \
    echo '{\"body\":\"...\",\"entities\":[],\"relationships\":[]}' | \\\n    \
    sqlite-graphrag remember --name my-mem --type note --description \"desc\" --graph-stdin\n\n  \
    # Enable GLiNER NER extraction with --graph-stdin\n  \
    echo '{\"body\":\"Alice from Microsoft...\",\"entities\":[],\"relationships\":[]}' | \\\n    \
    sqlite-graphrag remember --name ner-test --type note --description \"test\" \\\n    \
    --graph-stdin --enable-ner --gliner-variant int8\n\n  \
    # Idempotent upsert with --force-merge\n  \
    sqlite-graphrag remember --name my-mem --type note --description \"updated\" \\\n    \
    --body \"new content\" --force-merge\n\n\
NOTE:\n  \
    remember does NOT accept positional arguments.\n  \
    Use --body \"text\" for inline content\n  \
    Use --body-file path for file content\n  \
    Use --body-stdin for piped content\n  \
    Use --graph-stdin for JSON with entities and relationships\n\n\
ENTITY TYPES (for --graph-stdin entities, NOT memory --type):\n  \
    concept, tool, person, file, project, decision, incident,\n  \
    organization, location, date, dashboard, issue_tracker, memory\n  \
    WARNING: reference, skill, document, note, user, feedback are\n  \
    MEMORY types only — NOT valid for entities.\n  \
    Mapping: reference→concept, document→file, user→person")]
pub struct RememberArgs {
    /// Memory name in kebab-case (lowercase letters, digits, hyphens).
    /// Acts as unique key within the namespace; collisions trigger merge or rejection.
    #[arg(long)]
    pub name: String,
    #[arg(
        long,
        value_enum,
        long_help = "Memory kind stored in `memories.type`. Required when creating a new memory. Optional with --force-merge: if omitted the existing memory type is inherited. This is NOT the graph `entity_type` used in `--entities-file`. Valid values: user, feedback, project, reference, decision, incident, skill, document, note."
    )]
    pub r#type: Option<MemoryType>,
    /// Short description (≤500 chars) summarizing the memory for use in `list` and `recall` snippets.
    /// Required when creating a new memory. Optional with --force-merge: if omitted the existing description is inherited.
    #[arg(long)]
    pub description: Option<String>,
    /// Inline body content. Mutually exclusive with --body-file, --body-stdin, --graph-stdin.
    /// Maximum 512000 bytes; rejected if empty without an external graph.
    #[arg(
        long,
        help = "Inline body content (max 500 KB / 512000 bytes; for larger inputs split into multiple memories or use --body-file)",
        conflicts_with_all = ["body_file", "body_stdin", "graph_stdin"]
    )]
    pub body: Option<String>,
    #[arg(
        long,
        help = "Read body from a file instead of --body",
        conflicts_with_all = ["body", "body_stdin", "graph_stdin"]
    )]
    pub body_file: Option<std::path::PathBuf>,
    /// Read body from stdin until EOF. Useful in pipes (echo "..." | sqlite-graphrag remember ...).
    /// Mutually exclusive with --body, --body-file, --graph-stdin.
    #[arg(
        long,
        conflicts_with_all = ["body", "body_file", "graph_stdin"]
    )]
    pub body_stdin: bool,
    #[arg(
        long,
        help = "JSON file containing entities to associate with this memory"
    )]
    pub entities_file: Option<std::path::PathBuf>,
    #[arg(
        long,
        help = "JSON file containing relationships to associate with this memory"
    )]
    pub relationships_file: Option<std::path::PathBuf>,
    #[arg(
        long,
        help = "Read graph JSON (body + entities + relationships) from stdin",
        conflicts_with_all = [
            "body",
            "body_file",
            "body_stdin",
            "entities_file",
            "relationships_file"
        ]
    )]
    pub graph_stdin: bool,
    #[arg(
        long,
        help = "Namespace (env: SQLITE_GRAPHRAG_NAMESPACE, default: global)"
    )]
    pub namespace: Option<String>,
    /// Inline JSON object with arbitrary metadata key-value pairs. Mutually exclusive with --metadata-file.
    #[arg(long)]
    pub metadata: Option<String>,
    #[arg(long, help = "JSON file containing metadata key-value pairs")]
    pub metadata_file: Option<std::path::PathBuf>,
    #[arg(long)]
    pub force_merge: bool,
    #[arg(
        long,
        value_name = "EPOCH_OR_RFC3339",
        value_parser = crate::parsers::parse_expected_updated_at,
        long_help = "Optimistic lock: reject if updated_at does not match. \
Accepts Unix epoch (e.g. 1700000000) or RFC 3339 (e.g. 2026-04-19T12:00:00Z)."
    )]
    pub expected_updated_at: Option<i64>,
    #[arg(
        long,
        env = "SQLITE_GRAPHRAG_ENABLE_NER",
        value_parser = crate::parsers::parse_bool_flexible,
        action = clap::ArgAction::Set,
        num_args = 0..=1,
        default_missing_value = "true",
        default_value = "false",
        help = "Enable automatic GLiNER NER entity/relationship extraction from body"
    )]
    pub enable_ner: bool,
    #[arg(
        long,
        env = "SQLITE_GRAPHRAG_GLINER_VARIANT",
        default_value = "fp32",
        help = "GLiNER model variant: fp32 (1.1GB, best quality), fp16 (580MB), int8 (349MB, fastest but may miss entities on short texts), q4, q4f16"
    )]
    pub gliner_variant: String,
    #[arg(long, hide = true)]
    pub skip_extraction: bool,
    /// Explicitly clear the body content (set to empty string). Required to distinguish
    /// intentional body clearing from accidental omission during --force-merge.
    /// Without this flag, an empty body passed to --force-merge preserves the existing body.
    #[arg(
        long,
        default_value_t = false,
        help = "Explicitly clear body content during --force-merge (without this flag, an empty body is ignored and the existing body is kept)"
    )]
    pub clear_body: bool,
    /// Validate input and report planned actions without persisting.
    #[arg(
        long,
        default_value_t = false,
        help = "Validate input and report planned actions without persisting"
    )]
    pub dry_run: bool,
    /// Optional opaque session identifier for tracing memory provenance across multi-agent runs.
    #[arg(long)]
    pub session_id: Option<String>,
    #[arg(long, value_enum, default_value_t = JsonOutputFormat::Json)]
    pub format: JsonOutputFormat,
    #[arg(long, hide = true, help = "No-op; JSON is always emitted on stdout")]
    pub json: bool,
    #[arg(long, env = "SQLITE_GRAPHRAG_DB_PATH")]
    pub db: Option<String>,
    /// Maximum process RSS in MiB; abort if exceeded during embedding.
    #[arg(long, default_value_t = crate::constants::DEFAULT_MAX_RSS_MB,
          help = "Maximum process RSS in MiB; abort if exceeded during embedding (default: 8192)")]
    pub max_rss_mb: u64,
    /// Emit a warning (but do not reject) when persisting an entity whose degree would
    /// exceed this value after the upsert. Default 50. Set 0 to disable the check.
    #[arg(long, default_value_t = 50, value_name = "N")]
    pub max_entity_degree: u32,
}

#[derive(Deserialize, Default)]
#[serde(deny_unknown_fields)]
struct GraphInput {
    #[serde(default)]
    body: Option<String>,
    #[serde(default)]
    entities: Vec<NewEntity>,
    #[serde(default)]
    relationships: Vec<NewRelationship>,
}

fn normalize_and_validate_graph_input(graph: &mut GraphInput) -> Result<(), AppError> {
    for rel in &mut graph.relationships {
        rel.relation = crate::parsers::normalize_relation(&rel.relation);
        if let Err(e) = crate::parsers::validate_relation_format(&rel.relation) {
            return Err(AppError::Validation(format!(
                "{e} for relationship '{}' -> '{}'",
                rel.source, rel.target
            )));
        }
        crate::parsers::warn_if_non_canonical(&rel.relation);
        if !(0.0..=1.0).contains(&rel.strength) {
            return Err(AppError::Validation(format!(
                "invalid strength {} for relationship '{}' -> '{}'; expected value in [0.0, 1.0]",
                rel.strength, rel.source, rel.target
            )));
        }
    }

    Ok(())
}

#[tracing::instrument(skip_all, level = "debug", name = "remember")]
pub fn run(args: RememberArgs) -> Result<(), AppError> {
    use crate::constants::*;

    let inicio = std::time::Instant::now();
    let _ = args.format;
    tracing::debug!(target: "remember", name = %args.name, "persisting memory");
    let namespace = crate::namespace::resolve_namespace(args.namespace.as_deref())?;

    // Capture the original `--name` before normalization so the JSON response can
    // surface `name_was_normalized` + `original_name` (B_4 in v1.0.32). Stored as
    // an owned String because `args.name` is moved into the response below.
    let original_name = args.name.clone();

    // Auto-normalize to kebab-case before validation (P2-H).
    // v1.0.20: also trims hyphens at the boundary (including trailing) to avoid rejection
    // after truncation by a long filename ending in a hyphen.
    let normalized_name = {
        let lower = args.name.to_lowercase().replace(['_', ' '], "-");
        let trimmed = lower.trim_matches('-').to_string();
        if trimmed != args.name {
            tracing::warn!(target: "remember",
                original = %args.name,
                normalized = %trimmed,
                "name auto-normalized to kebab-case"
            );
        }
        trimmed
    };
    let name_was_normalized = normalized_name != original_name;

    if normalized_name.is_empty() {
        return Err(AppError::Validation(
            "name cannot be empty after normalization (input was blank or contained only hyphens/underscores/spaces)".to_string(),
        ));
    }
    if normalized_name.len() > MAX_MEMORY_NAME_LEN {
        return Err(AppError::LimitExceeded(
            crate::i18n::validation::name_length(MAX_MEMORY_NAME_LEN),
        ));
    }

    if normalized_name.starts_with("__") {
        return Err(AppError::Validation(
            crate::i18n::validation::reserved_name(),
        ));
    }

    {
        let slug_re = crate::constants::name_slug_regex();
        if !slug_re.is_match(&normalized_name) {
            return Err(AppError::Validation(crate::i18n::validation::name_kebab(
                &normalized_name,
            )));
        }
    }

    if let Some(ref desc) = args.description {
        if desc.len() > MAX_MEMORY_DESCRIPTION_LEN {
            return Err(AppError::Validation(
                crate::i18n::validation::description_exceeds(MAX_MEMORY_DESCRIPTION_LEN),
            ));
        }
    }

    let mut raw_body = if let Some(b) = args.body {
        b
    } else if let Some(ref path) = args.body_file {
        let file_size = std::fs::metadata(path).map_err(AppError::Io)?.len();
        if file_size > MAX_MEMORY_BODY_LEN as u64 {
            return Err(AppError::LimitExceeded(
                crate::i18n::validation::body_exceeds(MAX_MEMORY_BODY_LEN),
            ));
        }
        match std::fs::read_to_string(path) {
            Ok(s) => s,
            Err(e) if e.kind() == std::io::ErrorKind::InvalidData => {
                let bytes = std::fs::read(path).map_err(AppError::Io)?;
                tracing::warn!(target: "remember", "body file contains invalid UTF-8; replacing invalid sequences");
                String::from_utf8_lossy(&bytes).into_owned()
            }
            Err(e) => return Err(AppError::Io(e)),
        }
    } else if args.body_stdin || args.graph_stdin {
        crate::stdin_helper::read_stdin_with_timeout(60)?
    } else {
        String::new()
    };

    let mut entities_provided_externally =
        args.entities_file.is_some() || args.relationships_file.is_some();

    let mut graph = GraphInput::default();
    if let Some(path) = args.entities_file {
        let file_size = std::fs::metadata(&path).map_err(AppError::Io)?.len();
        if file_size > MAX_MEMORY_BODY_LEN as u64 {
            return Err(AppError::LimitExceeded(
                crate::i18n::validation::body_exceeds(MAX_MEMORY_BODY_LEN),
            ));
        }
        let content = std::fs::read_to_string(&path).map_err(AppError::Io)?;
        graph.entities = serde_json::from_str(&content)?;
    }
    if let Some(path) = args.relationships_file {
        let file_size = std::fs::metadata(&path).map_err(AppError::Io)?.len();
        if file_size > MAX_MEMORY_BODY_LEN as u64 {
            return Err(AppError::LimitExceeded(
                crate::i18n::validation::body_exceeds(MAX_MEMORY_BODY_LEN),
            ));
        }
        let content = std::fs::read_to_string(&path).map_err(AppError::Io)?;
        graph.relationships = serde_json::from_str(&content)?;
    }
    if args.graph_stdin {
        graph = serde_json::from_str::<GraphInput>(&raw_body).map_err(|e| {
            AppError::Validation(format!("invalid JSON payload on --graph-stdin: {e}"))
        })?;
        raw_body = graph.body.take().unwrap_or_default();
    }
    if args.graph_stdin && !graph.entities.is_empty() {
        entities_provided_externally = true;
    }

    if graph.entities.len() > max_entities_per_memory() {
        return Err(AppError::LimitExceeded(errors_msg::entity_limit_exceeded(
            max_entities_per_memory(),
        )));
    }
    let mut relationships_truncated = false;
    let rel_cap = max_relationships_per_memory();
    if graph.relationships.len() > rel_cap {
        tracing::warn!(target: "remember",
            count = graph.relationships.len(),
            cap = rel_cap,
            "truncating relationships to cap"
        );
        graph.relationships.truncate(rel_cap);
        relationships_truncated = true;
    }
    normalize_and_validate_graph_input(&mut graph)?;

    if raw_body.len() > MAX_MEMORY_BODY_LEN {
        return Err(AppError::LimitExceeded(
            crate::i18n::validation::body_exceeds(MAX_MEMORY_BODY_LEN),
        ));
    }

    // v1.0.22 P1: reject empty or whitespace-only body when no external graph is provided.
    // Without this check, empty embeddings would be persisted, breaking recall semantics.
    // GAP-08: skip this guard when --force-merge without --clear-body; the existing body
    // will be preserved from the database, so the effective body will not be empty.
    let body_will_be_preserved = args.force_merge && raw_body.trim().is_empty() && !args.clear_body;
    if !entities_provided_externally
        && graph.entities.is_empty()
        && raw_body.trim().is_empty()
        && !body_will_be_preserved
        && !args.clear_body
    {
        return Err(AppError::Validation(crate::i18n::validation::empty_body()));
    }

    let metadata: serde_json::Value = if let Some(m) = args.metadata {
        serde_json::from_str(&m)?
    } else if let Some(path) = args.metadata_file {
        let file_size = std::fs::metadata(&path).map_err(AppError::Io)?.len();
        if file_size > MAX_MEMORY_BODY_LEN as u64 {
            return Err(AppError::LimitExceeded(
                crate::i18n::validation::body_exceeds(MAX_MEMORY_BODY_LEN),
            ));
        }
        let content = std::fs::read_to_string(&path).map_err(AppError::Io)?;
        serde_json::from_str(&content)?
    } else {
        serde_json::json!({})
    };

    let mut body_hash = blake3::hash(raw_body.as_bytes()).to_hex().to_string();
    let mut snippet: String = raw_body.chars().take(200).collect();

    let paths = AppPaths::resolve(args.db.as_deref())?;
    paths.ensure_dirs()?;

    // v1.0.20: use .trim().is_empty() to reject bodies that are only whitespace.
    let mut extraction_method: Option<String> = None;
    let mut extracted_urls: Vec<crate::extraction::ExtractedUrl> = Vec::with_capacity(4);
    if args.enable_ner && args.skip_extraction {
        return Err(AppError::Validation(
            "--enable-ner and --skip-extraction are mutually exclusive; remove one".to_string(),
        ));
    }
    if args.skip_extraction && !args.enable_ner {
        return Err(AppError::Validation(
            "--skip-extraction is deprecated since v1.0.45 and has no effect; remove this flag"
                .to_string(),
        ));
    }
    let gliner_variant: crate::extraction::GlinerVariant =
        args.gliner_variant.parse().unwrap_or_else(|e| {
            tracing::warn!(target: "remember", error = %e, "invalid --gliner-variant, defaulting to fp32");
            crate::extraction::GlinerVariant::Fp32
        });
    if args.enable_ner && graph.entities.is_empty() && !raw_body.trim().is_empty() {
        match crate::extraction::extract_graph_auto(&raw_body, &paths, gliner_variant) {
            Ok(extracted) => {
                extraction_method = Some(extracted.extraction_method.clone());
                extracted_urls = extracted.urls;
                graph.entities = extracted.entities;
                graph.relationships = extracted.relationships;
                relationships_truncated = extracted.relationships_truncated;

                if graph.entities.len() > max_entities_per_memory() {
                    graph.entities.truncate(max_entities_per_memory());
                }
                if graph.relationships.len() > max_relationships_per_memory() {
                    relationships_truncated = true;
                    graph.relationships.truncate(max_relationships_per_memory());
                }
                normalize_and_validate_graph_input(&mut graph)?;
            }
            Err(e) => {
                tracing::warn!(target: "remember", error = %e, "auto-extraction failed, graceful degradation");
                extraction_method = Some("none:extraction-failed".to_string());
            }
        }
    }

    let mut conn = open_rw(&paths.db)?;
    ensure_schema(&mut conn)?;

    // --dry-run: emit planned action without any DB writes and return.
    if args.dry_run {
        let existing = memories::find_by_name(&conn, &namespace, &normalized_name)?;
        let planned_action = if existing.is_some() && args.force_merge {
            "would_update"
        } else {
            "would_create"
        };
        output::emit_json(&serde_json::json!({
            "dry_run": true,
            "name": normalized_name,
            "namespace": namespace,
            "planned_action": planned_action,
        }))?;
        return Ok(());
    }

    {
        use crate::constants::MAX_NAMESPACES_ACTIVE;
        let active_count: u32 = conn.query_row(
            "SELECT COUNT(DISTINCT namespace) FROM memories WHERE deleted_at IS NULL",
            [],
            |r| r.get::<_, i64>(0).map(|v| v as u32),
        )?;
        let ns_exists: bool = conn.query_row(
            "SELECT EXISTS(SELECT 1 FROM memories WHERE namespace = ?1 AND deleted_at IS NULL)",
            rusqlite::params![namespace],
            |r| r.get::<_, i64>(0).map(|v| v > 0),
        )?;
        if !ns_exists && active_count >= MAX_NAMESPACES_ACTIVE {
            return Err(AppError::NamespaceError(format!(
                "active namespace limit of {MAX_NAMESPACES_ACTIVE} reached while trying to create '{namespace}'"
            )));
        }
    }

    // M7: detect soft-deleted memory before the standard duplicate check.
    if let Some((sd_id, true)) =
        memories::find_by_name_any_state(&conn, &namespace, &normalized_name)?
    {
        if args.force_merge {
            memories::clear_deleted_at(&conn, sd_id)?;
        } else {
            return Err(AppError::Duplicate(
                errors_msg::duplicate_memory_soft_deleted(&normalized_name, &namespace),
            ));
        }
    }

    let existing_memory = memories::find_by_name(&conn, &namespace, &normalized_name)?;
    if existing_memory.is_some() && !args.force_merge {
        return Err(AppError::Duplicate(errors_msg::duplicate_memory(
            &normalized_name,
            &namespace,
        )));
    }

    // GAP-10: resolve type and description.
    // For CREATE path (new memory): both are required.
    // For UPDATE path (--force-merge on existing memory): inherit from existing row when omitted.
    let (resolved_type, resolved_description) = if existing_memory.is_none() {
        // CREATE path — both fields are mandatory.
        let t = args.r#type.ok_or_else(|| {
            AppError::Validation(
                "--type and --description are required when creating a new memory".to_string(),
            )
        })?;
        let d = args.description.clone().ok_or_else(|| {
            AppError::Validation(
                "--type and --description are required when creating a new memory".to_string(),
            )
        })?;
        (t.as_str().to_string(), d)
    } else {
        // UPDATE path (--force-merge) — inherit missing fields from stored row.
        let existing_row = memories::read_by_name(&conn, &namespace, &normalized_name)?
            .ok_or_else(|| {
                AppError::NotFound(format!(
                    "memory '{normalized_name}' not found in namespace '{namespace}'"
                ))
            })?;
        let t = args
            .r#type
            .map(|v| v.as_str().to_string())
            .unwrap_or_else(|| existing_row.memory_type.clone());
        let d = args
            .description
            .clone()
            .unwrap_or_else(|| existing_row.description.clone());
        (t, d)
    };

    // GAP-08/GAP-09: protect existing body from accidental destruction during --force-merge.
    // When the caller omits a body (or passes an empty one) without --clear-body, silently
    // preserve the existing body from the database.  This prevents a common scripting mistake
    // where a cron job updates metadata fields and inadvertently wipes the stored content.
    if body_will_be_preserved {
        if let Some(existing_row) = memories::read_by_name(&conn, &namespace, &normalized_name)? {
            if !existing_row.body.is_empty() {
                tracing::debug!(target: "remember",
                    name = %normalized_name,
                    "GAP-08: empty body with --force-merge and no --clear-body; preserving existing body"
                );
                raw_body = existing_row.body;
                body_hash = blake3::hash(raw_body.as_bytes()).to_hex().to_string();
                snippet = raw_body.chars().take(200).collect();
            }
        }
    }

    let duplicate_hash_id = memories::find_by_hash(&conn, &namespace, &body_hash)?;

    output::emit_progress_i18n(
        &format!(
            "Remember stage: validated input; available memory {} MB",
            crate::memory_guard::available_memory_mb()
        ),
        &format!(
            "Stage remember: input validated; available memory {} MB",
            crate::memory_guard::available_memory_mb()
        ),
    );

    let tokenizer = crate::tokenizer::get_tokenizer(&paths.models)?;
    let model_max_length = crate::tokenizer::get_model_max_length(&paths.models)?;
    let total_passage_tokens = crate::tokenizer::count_passage_tokens(tokenizer, &raw_body)?;
    let chunks_info = chunking::split_into_chunks_hierarchical(&raw_body, tokenizer);
    let chunks_created = chunks_info.len();
    // For single-chunk bodies the memory row itself stores the content and no
    // entry is appended to `memory_chunks` (see line ~545). For multi-chunk
    // bodies every chunk is persisted via `insert_chunk_slices`.
    let chunks_persisted = compute_chunks_persisted(chunks_info.len());

    output::emit_progress_i18n(
        &format!(
            "Remember stage: tokenizer counted {total_passage_tokens} passage tokens (model max {model_max_length}); chunking produced {} chunks; process RSS {} MB",
            chunks_created,
            crate::memory_guard::current_process_memory_mb().unwrap_or(0)
        ),
        &format!(
            "Stage remember: tokenizer counted {total_passage_tokens} passage tokens (model max {model_max_length}); chunking produced {} chunks; process RSS {} MB",
            chunks_created,
            crate::memory_guard::current_process_memory_mb().unwrap_or(0)
        ),
    );

    if chunks_created > crate::constants::REMEMBER_MAX_SAFE_MULTI_CHUNKS {
        return Err(AppError::LimitExceeded(format!(
            "document produces {chunks_created} chunks; current safe operational limit is {} chunks; split the document before using remember",
            crate::constants::REMEMBER_MAX_SAFE_MULTI_CHUNKS
        )));
    }

    output::emit_progress_i18n("Computing embedding...", "Calculando embedding...");
    let mut chunk_embeddings_cache: Option<Vec<Vec<f32>>> = None;

    let embedding = if chunks_info.len() == 1 {
        crate::daemon::embed_passage_or_local(&paths.models, &raw_body)?
    } else {
        let chunk_texts: Vec<&str> = chunks_info
            .iter()
            .map(|c| chunking::chunk_text(&raw_body, c))
            .collect();
        output::emit_progress_i18n(
            &format!(
                "Embedding {} chunks serially to keep memory bounded...",
                chunks_info.len()
            ),
            &format!(
                "Embedding {} chunks serially to keep memory bounded...",
                chunks_info.len()
            ),
        );
        let embed_cap = chunk_texts.len();
        let mut chunk_embeddings = Vec::new();
        chunk_embeddings.try_reserve(embed_cap).map_err(|_| {
            AppError::LimitExceeded(format!(
                "allocation of {embed_cap} chunk embeddings would exceed available memory"
            ))
        })?;
        for chunk_text in &chunk_texts {
            if let Some(rss) = crate::memory_guard::current_process_memory_mb() {
                if rss > args.max_rss_mb {
                    tracing::error!(target: "remember",
                        rss_mb = rss,
                        max_rss_mb = args.max_rss_mb,
                        "RSS exceeded --max-rss-mb threshold; aborting to prevent system instability"
                    );
                    return Err(AppError::LowMemory {
                        available_mb: crate::memory_guard::available_memory_mb(),
                        required_mb: args.max_rss_mb,
                    });
                }
            }
            chunk_embeddings.push(crate::daemon::embed_passage_or_local(
                &paths.models,
                chunk_text,
            )?);
        }
        output::emit_progress_i18n(
            &format!(
                "Remember stage: chunk embeddings complete; process RSS {} MB",
                crate::memory_guard::current_process_memory_mb().unwrap_or(0)
            ),
            &format!(
                "Stage remember: chunk embeddings completed; process RSS {} MB",
                crate::memory_guard::current_process_memory_mb().unwrap_or(0)
            ),
        );
        let aggregated = chunking::aggregate_embeddings(&chunk_embeddings);
        chunk_embeddings_cache = Some(chunk_embeddings);
        aggregated
    };
    let body_for_storage = raw_body;

    let memory_type = resolved_type.as_str();
    let new_memory = NewMemory {
        namespace: namespace.clone(),
        name: normalized_name.clone(),
        memory_type: memory_type.to_string(),
        description: resolved_description.clone(),
        body: body_for_storage,
        body_hash: body_hash.clone(),
        session_id: args.session_id.clone(),
        source: "agent".to_string(),
        metadata,
    };

    let mut warnings = Vec::with_capacity(4);
    let mut entities_persisted = 0usize;
    let mut relationships_persisted = 0usize;

    let graph_entity_embeddings = graph
        .entities
        .iter()
        .map(|entity| {
            let entity_text = match &entity.description {
                Some(desc) => format!("{} {}", entity.name, desc),
                None => entity.name.clone(),
            };
            crate::daemon::embed_passage_or_local(&paths.models, &entity_text)
        })
        .collect::<Result<Vec<_>, _>>()?;

    let tx = conn.transaction_with_behavior(rusqlite::TransactionBehavior::Immediate)?;

    let mut skip_reindex = false;
    let (memory_id, action, version) = match existing_memory {
        Some((existing_id, _updated_at, _current_version)) => {
            if let Some(hash_id) = duplicate_hash_id {
                if hash_id != existing_id {
                    warnings.push(format!(
                        "identical body already exists as memory id {hash_id}"
                    ));
                }
            }

            // C1 fix: capture old values for FTS5 sync before update
            let (old_fts_name, old_fts_desc, old_fts_body): (String, String, String) = tx
                .query_row(
                    "SELECT name, description, body FROM memories WHERE id = ?1",
                    rusqlite::params![existing_id],
                    |r| Ok((r.get(0)?, r.get(1)?, r.get(2)?)),
                )?;

            // G15: skip re-indexing when body hash matches (common in --force-merge loops)
            let existing_body_hash: Option<String> = tx
                .query_row(
                    "SELECT body_hash FROM memories WHERE id = ?1",
                    rusqlite::params![existing_id],
                    |r| r.get(0),
                )
                .ok();
            let body_unchanged = existing_body_hash.as_deref() == Some(&body_hash);
            skip_reindex = body_unchanged;
            if !body_unchanged {
                storage_chunks::delete_chunks(&tx, existing_id)?;
            }

            let next_v = versions::next_version(&tx, existing_id)?;
            memories::update(&tx, existing_id, &new_memory, args.expected_updated_at)?;

            // C1 fix: sync FTS5 external-content index after update
            // (trg_fts_au trigger is absent by design due to sqlite-vec conflict)
            memories::sync_fts_after_update(
                &tx,
                existing_id,
                &old_fts_name,
                &old_fts_desc,
                &old_fts_body,
                &normalized_name,
                &resolved_description,
                &new_memory.body,
            )?;

            versions::insert_version(
                &tx,
                existing_id,
                next_v,
                &normalized_name,
                memory_type,
                &resolved_description,
                &new_memory.body,
                &serde_json::to_string(&new_memory.metadata)?,
                None,
                "edit",
            )?;
            if !body_unchanged {
                memories::upsert_vec(
                    &tx,
                    existing_id,
                    &namespace,
                    memory_type,
                    &embedding,
                    &normalized_name,
                    &snippet,
                )?;
            }
            (existing_id, "updated".to_string(), next_v)
        }
        None => {
            if let Some(hash_id) = duplicate_hash_id {
                warnings.push(format!(
                    "identical body already exists as memory id {hash_id}"
                ));
            }
            let id = memories::insert(&tx, &new_memory)?;
            versions::insert_version(
                &tx,
                id,
                1,
                &normalized_name,
                memory_type,
                &resolved_description,
                &new_memory.body,
                &serde_json::to_string(&new_memory.metadata)?,
                None,
                "create",
            )?;
            memories::upsert_vec(
                &tx,
                id,
                &namespace,
                memory_type,
                &embedding,
                &normalized_name,
                &snippet,
            )?;
            (id, "created".to_string(), 1)
        }
    };

    if chunks_info.len() > 1 && !skip_reindex {
        storage_chunks::insert_chunk_slices(&tx, memory_id, &new_memory.body, &chunks_info)?;

        let chunk_embeddings = chunk_embeddings_cache.take().ok_or_else(|| {
            AppError::Internal(anyhow::anyhow!(
                "chunk embeddings cache missing in multi-chunk remember path"
            ))
        })?;

        for (i, emb) in chunk_embeddings.iter().enumerate() {
            storage_chunks::upsert_chunk_vec(&tx, i as i64, memory_id, i as i32, emb)?;
        }
        output::emit_progress_i18n(
            &format!(
                "Remember stage: persisted chunk vectors; process RSS {} MB",
                crate::memory_guard::current_process_memory_mb().unwrap_or(0)
            ),
            &format!(
                "Etapa remember: vetores de chunks persistidos; RSS do processo {} MB",
                crate::memory_guard::current_process_memory_mb().unwrap_or(0)
            ),
        );
    }

    if !graph.entities.is_empty() || !graph.relationships.is_empty() {
        for entity in &graph.entities {
            let entity_id = entities::upsert_entity(&tx, &namespace, entity)?;
            let entity_embedding = &graph_entity_embeddings[entities_persisted];
            entities::upsert_entity_vec(
                &tx,
                entity_id,
                &namespace,
                entity.entity_type,
                entity_embedding,
                &entity.name,
            )?;
            entities::link_memory_entity(&tx, memory_id, entity_id)?;
            entities::increment_degree(&tx, entity_id)?;
            // GAP-17: warn when entity degree exceeds the configured cap.
            if args.max_entity_degree > 0 {
                let cap = args.max_entity_degree as i64;
                let degree: i64 = tx.query_row(
                    "SELECT degree FROM entities WHERE id = ?1",
                    rusqlite::params![entity_id],
                    |r| r.get(0),
                )?;
                if degree > cap {
                    tracing::warn!(target: "remember",
                        entity = %entity.name,
                        degree = degree,
                        cap = cap,
                        "entity degree cap exceeded"
                    );
                }
            }
            entities_persisted += 1;
        }
        let entity_types: std::collections::HashMap<&str, EntityType> = graph
            .entities
            .iter()
            .map(|entity| (entity.name.as_str(), entity.entity_type))
            .collect();

        for rel in &graph.relationships {
            let source_entity = NewEntity {
                name: rel.source.clone(),
                entity_type: entity_types
                    .get(rel.source.as_str())
                    .copied()
                    .unwrap_or(EntityType::Concept),
                description: None,
            };
            let target_entity = NewEntity {
                name: rel.target.clone(),
                entity_type: entity_types
                    .get(rel.target.as_str())
                    .copied()
                    .unwrap_or(EntityType::Concept),
                description: None,
            };
            let source_id = entities::upsert_entity(&tx, &namespace, &source_entity)?;
            let target_id = entities::upsert_entity(&tx, &namespace, &target_entity)?;
            let rel_id = entities::upsert_relationship(&tx, &namespace, source_id, target_id, rel)?;
            entities::link_memory_relationship(&tx, memory_id, rel_id)?;
            relationships_persisted += 1;
        }
    }
    tx.commit()?;

    // v1.0.24 P0-2: persist URLs in a dedicated table, outside the main transaction.
    // Failures do not propagate — non-critical path with graceful degradation.
    let urls_persisted = if !extracted_urls.is_empty() {
        let url_entries: Vec<storage_urls::MemoryUrl> = extracted_urls
            .into_iter()
            .map(|u| storage_urls::MemoryUrl {
                url: u.url,
                offset: Some(u.offset as i64),
            })
            .collect();
        storage_urls::insert_urls(&conn, memory_id, &url_entries)
    } else {
        0
    };

    conn.execute_batch("PRAGMA wal_checkpoint(TRUNCATE);")?;

    let created_at_epoch = chrono::Utc::now().timestamp();
    let created_at_iso = crate::tz::format_iso(chrono::Utc::now());

    output::emit_json(&RememberResponse {
        memory_id,
        // Persist the normalized (kebab-case) slug as `name` since that is the
        // storage key. The original input is exposed via `original_name` only
        // when normalization actually changed something (B_4 in v1.0.32).
        name: normalized_name.clone(),
        namespace,
        action: action.clone(),
        operation: action,
        version,
        entities_persisted,
        relationships_persisted,
        relationships_truncated,
        chunks_created,
        chunks_persisted,
        urls_persisted,
        extraction_method,
        merged_into_memory_id: None,
        warnings,
        created_at: created_at_epoch,
        created_at_iso,
        elapsed_ms: inicio.elapsed().as_millis() as u64,
        name_was_normalized,
        original_name: name_was_normalized.then_some(original_name),
    })?;

    Ok(())
}

#[cfg(test)]
mod tests {
    use super::compute_chunks_persisted;
    use crate::output::RememberResponse;

    // Bug H-M8: chunks_persisted contract is unit-testable and matches schema.
    #[test]
    fn chunks_persisted_zero_for_zero_chunks() {
        assert_eq!(compute_chunks_persisted(0), 0);
    }

    #[test]
    fn chunks_persisted_zero_for_single_chunk_body() {
        // Single-chunk bodies live in the memories row itself; no row is
        // appended to memory_chunks. This is the documented contract.
        assert_eq!(compute_chunks_persisted(1), 0);
    }

    #[test]
    fn chunks_persisted_equals_count_for_multi_chunk_body() {
        // Every chunk above the first triggers a row in memory_chunks.
        assert_eq!(compute_chunks_persisted(2), 2);
        assert_eq!(compute_chunks_persisted(7), 7);
        assert_eq!(compute_chunks_persisted(64), 64);
    }

    #[test]
    fn remember_response_serializes_required_fields() {
        let resp = RememberResponse {
            memory_id: 42,
            name: "minha-mem".to_string(),
            namespace: "global".to_string(),
            action: "created".to_string(),
            operation: "created".to_string(),
            version: 1,
            entities_persisted: 0,
            relationships_persisted: 0,
            relationships_truncated: false,
            chunks_created: 1,
            chunks_persisted: 0,
            urls_persisted: 0,
            extraction_method: None,
            merged_into_memory_id: None,
            warnings: vec![],
            created_at: 1_705_320_000,
            created_at_iso: "2024-01-15T12:00:00Z".to_string(),
            elapsed_ms: 55,
            name_was_normalized: false,
            original_name: None,
        };

        let json = serde_json::to_value(&resp).expect("serialization failed");
        assert_eq!(json["memory_id"], 42);
        assert_eq!(json["action"], "created");
        assert_eq!(json["operation"], "created");
        assert_eq!(json["version"], 1);
        assert_eq!(json["elapsed_ms"], 55u64);
        assert!(json["warnings"].is_array());
        assert!(json["merged_into_memory_id"].is_null());
    }

    #[test]
    fn remember_response_action_e_operation_sao_aliases() {
        let resp = RememberResponse {
            memory_id: 1,
            name: "mem".to_string(),
            namespace: "global".to_string(),
            action: "updated".to_string(),
            operation: "updated".to_string(),
            version: 2,
            entities_persisted: 3,
            relationships_persisted: 1,
            relationships_truncated: false,
            extraction_method: None,
            chunks_created: 2,
            chunks_persisted: 2,
            urls_persisted: 0,
            merged_into_memory_id: None,
            warnings: vec![],
            created_at: 0,
            created_at_iso: "1970-01-01T00:00:00Z".to_string(),
            elapsed_ms: 0,
            name_was_normalized: false,
            original_name: None,
        };

        let json = serde_json::to_value(&resp).expect("serialization failed");
        assert_eq!(
            json["action"], json["operation"],
            "action e operation devem ser iguais"
        );
        assert_eq!(json["entities_persisted"], 3);
        assert_eq!(json["relationships_persisted"], 1);
        assert_eq!(json["chunks_created"], 2);
    }

    #[test]
    fn remember_response_warnings_lista_mensagens() {
        let resp = RememberResponse {
            memory_id: 5,
            name: "dup-mem".to_string(),
            namespace: "global".to_string(),
            action: "created".to_string(),
            operation: "created".to_string(),
            version: 1,
            entities_persisted: 0,
            extraction_method: None,
            relationships_persisted: 0,
            relationships_truncated: false,
            chunks_created: 1,
            chunks_persisted: 0,
            urls_persisted: 0,
            merged_into_memory_id: None,
            warnings: vec!["identical body already exists as memory id 3".to_string()],
            created_at: 0,
            created_at_iso: "1970-01-01T00:00:00Z".to_string(),
            elapsed_ms: 10,
            name_was_normalized: false,
            original_name: None,
        };

        let json = serde_json::to_value(&resp).expect("serialization failed");
        let warnings = json["warnings"]
            .as_array()
            .expect("warnings deve ser array");
        assert_eq!(warnings.len(), 1);
        assert!(warnings[0].as_str().unwrap().contains("identical body"));
    }

    #[test]
    fn invalid_name_reserved_prefix_returns_validation_error() {
        use crate::errors::AppError;
        // Validates the rejection logic for names with the "__" prefix directly
        let nome = "__reservado";
        let resultado: Result<(), AppError> = if nome.starts_with("__") {
            Err(AppError::Validation(
                crate::i18n::validation::reserved_name(),
            ))
        } else {
            Ok(())
        };
        assert!(resultado.is_err());
        if let Err(AppError::Validation(msg)) = resultado {
            assert!(!msg.is_empty());
        }
    }

    #[test]
    fn name_too_long_returns_validation_error() {
        use crate::errors::AppError;
        let nome_longo = "a".repeat(crate::constants::MAX_MEMORY_NAME_LEN + 1);
        let resultado: Result<(), AppError> =
            if nome_longo.is_empty() || nome_longo.len() > crate::constants::MAX_MEMORY_NAME_LEN {
                Err(AppError::Validation(crate::i18n::validation::name_length(
                    crate::constants::MAX_MEMORY_NAME_LEN,
                )))
            } else {
                Ok(())
            };
        assert!(resultado.is_err());
    }

    #[test]
    fn remember_response_merged_into_memory_id_some_serializes_integer() {
        let resp = RememberResponse {
            memory_id: 10,
            name: "mem-mergeada".to_string(),
            namespace: "global".to_string(),
            action: "updated".to_string(),
            operation: "updated".to_string(),
            version: 3,
            extraction_method: None,
            entities_persisted: 0,
            relationships_persisted: 0,
            relationships_truncated: false,
            chunks_created: 1,
            chunks_persisted: 0,
            urls_persisted: 0,
            merged_into_memory_id: Some(7),
            warnings: vec![],
            created_at: 0,
            created_at_iso: "1970-01-01T00:00:00Z".to_string(),
            elapsed_ms: 0,
            name_was_normalized: false,
            original_name: None,
        };

        let json = serde_json::to_value(&resp).expect("serialization failed");
        assert_eq!(json["merged_into_memory_id"], 7);
    }

    #[test]
    fn remember_response_urls_persisted_serializes_field() {
        // v1.0.24 P0-2: garante que urls_persisted aparece no JSON e aceita valor > 0.
        let resp = RememberResponse {
            memory_id: 3,
            name: "mem-com-urls".to_string(),
            namespace: "global".to_string(),
            action: "created".to_string(),
            operation: "created".to_string(),
            version: 1,
            entities_persisted: 0,
            relationships_persisted: 0,
            relationships_truncated: false,
            chunks_created: 1,
            chunks_persisted: 0,
            urls_persisted: 3,
            extraction_method: Some("regex-only".to_string()),
            merged_into_memory_id: None,
            warnings: vec![],
            created_at: 0,
            created_at_iso: "1970-01-01T00:00:00Z".to_string(),
            elapsed_ms: 0,
            name_was_normalized: false,
            original_name: None,
        };
        let json = serde_json::to_value(&resp).expect("serialization failed");
        assert_eq!(json["urls_persisted"], 3);
    }

    #[test]
    fn empty_name_after_normalization_returns_specific_message() {
        // P0-4 regression: name consisting only of hyphens normalizes to empty string;
        // must produce a distinct error message, not the "too long" message.
        use crate::errors::AppError;
        let normalized = "---".to_lowercase().replace(['_', ' '], "-");
        let normalized = normalized.trim_matches('-').to_string();
        let resultado: Result<(), AppError> = if normalized.is_empty() {
            Err(AppError::Validation(
                "name cannot be empty after normalization (input was blank or contained only hyphens/underscores/spaces)".to_string(),
            ))
        } else {
            Ok(())
        };
        assert!(resultado.is_err());
        if let Err(AppError::Validation(msg)) = resultado {
            assert!(
                msg.contains("empty after normalization"),
                "mensagem deve mencionar 'empty after normalization', obteve: {msg}"
            );
        }
    }

    #[test]
    fn name_only_underscores_after_normalization_returns_specific_message() {
        // P0-4 regression: name consisting only of underscores normalizes to empty string.
        use crate::errors::AppError;
        let normalized = "___".to_lowercase().replace(['_', ' '], "-");
        let normalized = normalized.trim_matches('-').to_string();
        assert!(
            normalized.is_empty(),
            "underscores devem normalizar para string vazia"
        );
        let resultado: Result<(), AppError> = if normalized.is_empty() {
            Err(AppError::Validation(
                "name cannot be empty after normalization (input was blank or contained only hyphens/underscores/spaces)".to_string(),
            ))
        } else {
            Ok(())
        };
        assert!(resultado.is_err());
        if let Err(AppError::Validation(msg)) = resultado {
            assert!(
                msg.contains("empty after normalization"),
                "mensagem deve mencionar 'empty after normalization', obteve: {msg}"
            );
        }
    }

    #[test]
    fn remember_response_relationships_truncated_serializes_field() {
        // P1-D: garante que relationships_truncated aparece no JSON como bool.
        let resp_false = RememberResponse {
            memory_id: 1,
            name: "test".to_string(),
            namespace: "global".to_string(),
            action: "created".to_string(),
            operation: "created".to_string(),
            version: 1,
            entities_persisted: 2,
            relationships_persisted: 1,
            relationships_truncated: false,
            chunks_created: 1,
            chunks_persisted: 0,
            urls_persisted: 0,
            extraction_method: None,
            merged_into_memory_id: None,
            warnings: vec![],
            created_at: 0,
            created_at_iso: "1970-01-01T00:00:00Z".to_string(),
            elapsed_ms: 0,
            name_was_normalized: false,
            original_name: None,
        };
        let json_false = serde_json::to_value(&resp_false).expect("serialization failed");
        assert_eq!(json_false["relationships_truncated"], false);

        let resp_true = RememberResponse {
            relationships_truncated: true,
            ..resp_false
        };
        let json_true = serde_json::to_value(&resp_true).expect("serialization failed");
        assert_eq!(json_true["relationships_truncated"], true);
    }

    // GAP-08: body-preservation predicate tests.
    // Verifies the decision logic that determines whether an existing body should
    // be kept instead of overwritten with an empty incoming body during --force-merge.

    /// Returns `true` when the existing body should be preserved.
    ///
    /// Mirrors the `body_will_be_preserved` expression in `run()` so the logic
    /// is testable without a real database connection.
    fn should_preserve_body(force_merge: bool, raw_body_is_empty: bool, clear_body: bool) -> bool {
        force_merge && raw_body_is_empty && !clear_body
    }

    #[test]
    fn gap08_empty_body_force_merge_no_clear_body_preserves() {
        // Caller passes no body with --force-merge but without --clear-body.
        // The existing body in the DB must be kept.
        assert!(
            should_preserve_body(true, true, false),
            "empty body + force-merge + no clear-body should trigger preservation"
        );
    }

    #[test]
    fn gap08_empty_body_force_merge_with_clear_body_does_not_preserve() {
        // Caller explicitly passes --clear-body; intentional wipe is honoured.
        assert!(
            !should_preserve_body(true, true, true),
            "--clear-body must bypass preservation"
        );
    }

    #[test]
    fn gap08_non_empty_body_force_merge_does_not_preserve() {
        // Caller provides a real body; it must overwrite the existing one.
        assert!(
            !should_preserve_body(true, false, false),
            "non-empty body must overwrite, not preserve"
        );
    }

    #[test]
    fn gap08_empty_body_no_force_merge_does_not_preserve() {
        // Without --force-merge the path is a fresh create; no preservation needed.
        assert!(
            !should_preserve_body(false, true, false),
            "no --force-merge means no preservation logic applies"
        );
    }
}