memleaf 0.2.53__tar.gz → 0.2.55__tar.gz

{memleaf-0.2.53 → memleaf-0.2.55}/CHANGELOG.md

@@ -2,6 +2,22 @@
 
 All notable changes to memleaf are documented here.
 
+## 0.2.55 — 2026-09-14
+
+- Classify memory candidates by future reuse with a shorter B3 instruction: point-in-time counts and snapshots without a trend, threshold, obligation, decision or later comparison are `no_future_value`, while future-use facts remain eligible. Project ownership now has a conservative Core guard: an explicit single-project label cannot silently persist as `global`, but ordinary platform, product and vendor mentions still do not establish ownership. Self-contained memories refer to the conversation person as the user rather than as an owner. The B3 system prompt shrinks from 6395 to 5686 UTF-8 bytes.
+- Preserve candidate-local date authority while improving structured todo deadlines. A legal `whole_unit` claim continues through the existing binding validator as one canonical full-text quote; no second evidence path was added. Core fills `due_date` only when that candidate's own admitted user/assistant evidence establishes one unambiguous deadline, never from sibling evidence, and exposes ambiguous deadline counts in total, stage and operation metrics instead of guessing.
+- Normalize compact date-time text without corrupting adjacent content: a resolved ISO date next to a valid clock gains one separating space, existing spacing remains unchanged, and non-date words such as `本周日报` remain ordinary text. Deadline recognition uses boundary-safe English cues, accepts slash dates before sentence punctuation, and rejects word-suffix and decimal false positives.
+
+Verification for this release used eight local synthetic regression tests with no model call and no production Vault write. They cover whole-unit/exact-quote equivalence, candidate date isolation, date-time boundaries, future-value protocol outcomes, explicit-project/global conflicts, platform-only mentions, unique and ambiguous todo deadlines, deadline word boundaries and metric visibility. Python 3.11 compilation and imports pass, as does `git diff --check`. The repository intentionally ignores `/tests/`, so these local tests are not part of the source distribution; real-model adherence and a newly installed Hermes replay remain post-release runtime acceptance.
+
+## 0.2.54 — 2026-09-14
+
+- Judge automatic memories by future reuse rather than speaker identity. Visible assistant final reports may establish facts learned from external sources, while transient tool failures, one-time fallbacks and routine checks with no issue or follow-up are `no_future_value`. The B3 system prompt and compact contract were rewritten rather than extended: together they shrink from 9469 to 9331 bytes, and the normal path still uses one model request.
+- Reject an unrelated `NO_CHANGE` target before it can enter durable audit state. B3 previously verified only that the selected ID existed in the local catalog; it now compares the candidate's exact cited span with the target title/body using source-neutral lexical anchors. An unproven relation defers only that candidate as `target_ambiguous`, without adding a model call or failing valid siblings. Long multi-topic whole-unit claims are likewise too broad to authorize `NO_CHANGE`.
+- Keep dates candidate-local and stop treating `本周日报` as `本周日` plus `报`. Date grounding now uses the candidate's exact validated quote and matching event timestamp instead of the whole assistant reply, so one mail item's date cannot authorize another memory. The relative-date tokenizer preserves `本周日报` and `下周日报`, while the existing `本周日` and `下周日` conversions remain intact.
+
+Verification for this release used a temporary focused regression module and no model call or production Vault write. Nine tests pass for unrelated and related `NO_CHANGE` targets, Chinese subjects, identifiers and numeric codes, long whole-unit deferral, assistant-reported future-use todos, execution-noise disposition, candidate-local date grounding, and the `本周日报`/`本周日` boundary. The three changed modules compile under Python 3.11, `git diff --check` passes, and no real-model adherence or installed-Hermes replay is claimed.
+
 ## 0.2.53 — 2026-09-13
 
 - Keep the memory when only the project name is unproven. A real turn was lost in full because of one word: the user wrote "在弄个记账的小玩意儿，就扔家里那台 N100 上跑，没打算上云", the model named the project `project:记账小玩意儿` -- the user's phrase with 的 removed -- and Core refuses a project scope whose name is not in the candidate's own evidence. That refusal discarded the whole candidate, so the N100 deployment and the no-cloud decision went with the unproven name, leaving three evidence units unresolved and nothing written. The memory was grounded; only the affiliation was not. The claim is now dropped and the memory kept with `global`, which asserts no ownership and loses no source-backed content. The refusal itself is unchanged: Core still never records a project it cannot ground.

{memleaf-0.2.53/src/memleaf.egg-info → memleaf-0.2.55}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: memleaf
-Version: 0.2.53
+Version: 0.2.55
 Summary: A local-first Markdown memory core for AI agents
 Author: memleaf contributors
 License-Expression: MIT
@@ -23,7 +23,7 @@ Dynamic: license-file
 
 [English](README.en.md) · [PyPI](https://pypi.org/project/memleaf/) · [GitHub](https://github.com/miffyblueboo/memleaf)
 
-> **版本：0.2.53。**
+> **版本：0.2.55。**
 > 记忆只从用户与 Agent 的可见对话提炼。Agent 已在回复中整理的事实、项目进展和明确待办可作为来源；邮件、附件、网页和终端等工具原文不进入记忆提炼。模型负责保留原话中的不确定性，现有记忆仅用于比较、去重和更新。Markdown 仍是唯一事实源。
 > **当前版本支持 Hermes 和 Codex。** Antigravity（反重力）不检测、不安装、不配置。
 

{memleaf-0.2.53 → memleaf-0.2.55}/README.en.md

@@ -4,7 +4,7 @@
 
 [中文](README.md) · [PyPI](https://pypi.org/project/memleaf/) · [GitHub](https://github.com/miffyblueboo/memleaf)
 
-> **Version: 0.2.53.**
+> **Version: 0.2.55.**
 > Automatic extraction now uses only the current turn's visible user input and final assistant reply. Raw tool output, attachments, web/file/terminal payloads and legacy tool-evidence bodies are not new source evidence; existing or retrieved memory remains comparison context rather than source authority. Background processing is persisted as a local job and can be checked through the read-only `process_status` MCP tool; failed work remains retryable and fail closed. The release also tightens source/date grounding, target reconciliation, duplicate/no-op handling and semantic review before writes. Markdown remains the sole source of truth with no SQLite runtime dependency. Acceptance covers deterministic regression suites and synthetic inputs; it does not claim real-mail or customer-business acceptance.
 > **The current release supports Hermes and Codex.** Antigravity is not detected, installed, or configured.
 

{memleaf-0.2.53 → memleaf-0.2.55}/README.md

@@ -4,7 +4,7 @@
 
 [English](README.en.md) · [PyPI](https://pypi.org/project/memleaf/) · [GitHub](https://github.com/miffyblueboo/memleaf)
 
-> **版本：0.2.53。**
+> **版本：0.2.55。**
 > 记忆只从用户与 Agent 的可见对话提炼。Agent 已在回复中整理的事实、项目进展和明确待办可作为来源；邮件、附件、网页和终端等工具原文不进入记忆提炼。模型负责保留原话中的不确定性，现有记忆仅用于比较、去重和更新。Markdown 仍是唯一事实源。
 > **当前版本支持 Hermes 和 Codex。** Antigravity（反重力）不检测、不安装、不配置。
 

{memleaf-0.2.53 → memleaf-0.2.55}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "memleaf"
-version = "0.2.53"
+version = "0.2.55"
 description = "A local-first Markdown memory core for AI agents"
 readme = "README.md"
 requires-python = ">=3.11"

{memleaf-0.2.53 → memleaf-0.2.55}/src/memleaf/__init__.py

@@ -1,6 +1,6 @@
 """Local-first Markdown memory core for AI agents."""
 
-__version__ = "0.2.53"
+__version__ = "0.2.55"
 
 from .config import DEFAULT_CONFIG, default_config, load_config, save_config
 from .frontmatter import FrontmatterError, dump_frontmatter, dump_yaml, load_yaml, parse_frontmatter

{memleaf-0.2.53 → memleaf-0.2.55}/src/memleaf/hermes_provider/plugin.yaml

@@ -1,5 +1,5 @@
 name: memleaf
-version: 0.2.53
+version: 0.2.55
 description: "Hermes-native external MemoryProvider for local-first Markdown memory shared with the memleaf MCP server."
 hooks:
   - prefetch

{memleaf-0.2.53 → memleaf-0.2.55}/src/memleaf/model_execution.py

@@ -56,7 +56,7 @@ _METRIC_EVENT_FIELDS = frozenset({
     # first two were emitted but silently dropped here, which made the
     # "visible rather than silent" claim in their release notes untrue.
     "b3_normalization_count", "b3_candidate_deferred_count",
-    "b3_ungrounded_scope_dropped_count",
+    "b3_ungrounded_scope_dropped_count", "b3_due_date_ambiguous_count",
 })
 _THINKING_OBSERVATION_SOURCES = frozenset({
     "request_parameter", "reasoning_tokens", "reasoning_content", "unavailable",
@@ -91,6 +91,7 @@ def _metric_bucket() -> dict[str, Any]:
         "b3_normalization_count": 0,
         "b3_candidate_deferred_count": 0,
         "b3_ungrounded_scope_dropped_count": 0,
+        "b3_due_date_ambiguous_count": 0,
         "_first_started": None,
         "_last_finished": None,
     }
@@ -515,6 +516,7 @@ class ModelExecutor:
             "repair_rejected_semantic_drift_count": int(bucket.get("repair_rejected_semantic_drift_count", 0)),
             "parse_accepted_count": int(bucket.get("parse_accepted_count", 0)),
             "decision_case_normalization_count": int(bucket.get("decision_case_normalization_count", 0)),
+            "b3_due_date_ambiguous_count": int(bucket.get("b3_due_date_ambiguous_count", 0)),
         }
         for field in _PROVIDER_METRIC_FIELDS:
             result[field] = int(bucket.get(field, 0))

{memleaf-0.2.53 → memleaf-0.2.55}/src/memleaf/process_common.py

@@ -16,6 +16,7 @@ from .llm import MODEL_ERROR_CODES, MODEL_VALIDATION_REASONS, ModelUnavailable
 from .locking import read_json
 from .models import Memory, utc_now
 from .retrieval import candidate_matches_query, normalize_term
+from .scope_state import project_scope_matches_text
 from .validation import MODEL_EVIDENCE_CHECKS, MODEL_VALIDATION_DETAILS, ModelOutputError, parse_strict_json, normalize_relative_calendar_text, safe_evidence_context, safe_schema_context
 
 _PROCESSING_LEASE_SECONDS = 3600
@@ -73,6 +74,26 @@ _SUMMARY_CHINESE_CALENDAR_DATE_RE = re.compile(
     r"(?![A-Za-z\d])"
 )
 
+_SLASH_CALENDAR_DATE_RE = re.compile(
+    r"(?<![\d./-])(?:(?P<year>\d{4})/)?(?P<month>\d{1,2})/(?P<day>\d{1,2})(?![\d/-]|\.\d)"
+)
+_DEADLINE_PREFIX_RE = re.compile(
+    r"(?:截止(?:日期|时间)?|截至(?:日期|时间)?|不晚于|不得晚于|\bdeadline\b|\bdue(?:\s+date)?\b|\bby\b|\bbefore\b)"
+    r"\s*(?:为|是|[:：])?\s*$",
+    re.IGNORECASE,
+)
+_DEADLINE_SUFFIX_RE = re.compile(
+    r"^\s*(?:(?:[01]?\d|2[0-3]):[0-5]\d\s*)?(?:前|之前|为止|截止|截至|不晚于|不得晚于|before\b|deadline\b|due\b)",
+    re.IGNORECASE,
+)
+_EXPLICIT_PROJECT_LABEL_RE = re.compile(
+    r"(?<![A-Za-z0-9_])project\s*:\s*(?P<colon>[^\s,，;；。！？:：]{1,64})"
+    r"|(?<![A-Za-z0-9_])project\s+(?P<english>[^\s,，;；。！？:：]{1,64})\s*[:：—–-]"
+    r"|(?:^|[\s(（\[{,，])项目\s*[:：]\s*(?P<chinese>[^\s,，;；。！？:：]{1,64})"
+    r"|(?:^|[\s(（\[{,，])(?P<suffix>[^\s,，;；。！？:：]{1,64})\s*项目\s*[:：]",
+    re.IGNORECASE,
+)
+
 
 
 
@@ -141,6 +162,71 @@ def _project_scope_occurrences(
     return selected
 
 
+def _explicit_project_scope_labels(
+    texts: Iterable[Any],
+    scope_registry: Mapping[str, Any] | None,
+) -> list[str]:
+    """Find project scopes explicitly written as project labels in candidate text.
+
+    A registered project name appearing as an ordinary noun is insufficient;
+    this recognizes only explicit forms such as ``project:Alpha`` or
+    ``Alpha 项目``. It supports a conservative global-scope guard without
+    treating a platform, product, or vendor mention as ownership.
+    """
+
+    rows = [text for text in texts if isinstance(text, str) and text.strip()]
+    if not rows:
+        return []
+    registry = scope_registry if isinstance(scope_registry, Mapping) else {}
+    try:
+        matches = {
+            scope
+            for text in rows
+            for scope in project_scope_matches_text(text, {"scopes": registry})
+        }
+    except (TypeError, ValueError):
+        return []
+    labeled: set[str] = set()
+    for scope in matches:
+        node = registry.get(scope)
+        terms = [scope.partition(":")[2]]
+        if isinstance(node, Mapping) and isinstance(node.get("aliases"), list):
+            terms.extend(value for value in node["aliases"] if isinstance(value, str))
+        for text in rows:
+            for raw_term in terms:
+                term = re.sub(r"\\\s+", r"\\s+", re.escape(raw_term.strip()))
+                if not term:
+                    continue
+                patterns = (
+                    rf"(?<![A-Za-z0-9_])project\s*:\s*{term}(?![A-Za-z0-9_])",
+                    rf"(?<![A-Za-z0-9_])project\s+{term}(?![A-Za-z0-9_])",
+                    rf"(?:^|[\s(（\[,，:：])项目\s*[:：]?\s*{term}(?=$|[\s，。,:：\-—])",
+                    rf"(?:^|[\s(（\[,，:：]){term}\s*(?:项目|project)(?=$|[\s，。,:：\-—])",
+                )
+                if any(re.search(pattern, text, re.IGNORECASE) for pattern in patterns):
+                    labeled.add(scope)
+                    break
+            if scope in labeled:
+                break
+    for text in rows:
+        for match in _EXPLICIT_PROJECT_LABEL_RE.finditer(text):
+            raw_name = next((value for value in match.groupdict().values() if value), "")
+            if not raw_name:
+                continue
+            try:
+                resolved = project_scope_matches_text(raw_name, {"scopes": registry})
+            except (TypeError, ValueError):
+                resolved = []
+            if len(resolved) == 1:
+                labeled.add(resolved[0])
+            else:
+                # An explicit but unregistered project label is still enough
+                # to reject a silent global attribution; the candidate is
+                # deferred because Core must not invent the project's scope.
+                labeled.add("project-label:" + raw_name.casefold())
+    return sorted(labeled, key=str.casefold)
+
+
 def _automatic_read_only_query(events: Iterable[Mapping[str, Any]]) -> bool:
     return read_only_turn(analyze_turn_evidence(events))
 
@@ -867,6 +953,72 @@ def _grounded_due_dates(
     return result
 
 
+def _grounded_deadline_dates(
+    evidence_events: Iterable[Mapping[str, Any]],
+) -> set[str]:
+    """Return only dates explicitly tied to a deadline in this evidence set.
+
+    Relative dates use the cited event's own timestamp. Yearless calendar dates
+    use that same timestamp only when one is available. No other turn content
+    or retrieval timestamp participates.
+    """
+
+    result: set[str] = set()
+    for event in evidence_events:
+        if not isinstance(event, Mapping) or event.get("role") not in {"user", "assistant"}:
+            continue
+        content = event.get("content")
+        if not isinstance(content, str) or not content:
+            continue
+        timestamp = _parse_time(event.get("timestamp"))
+        text = normalize_relative_calendar_text(content, timestamp) if timestamp is not None else content
+        if text is None:
+            text = content
+        tokens: list[tuple[int, int, str]] = []
+
+        for match in _ISO_CALENDAR_DATE_RE.finditer(text):
+            try:
+                canonical = datetime.strptime(match.group(1), "%Y-%m-%d").date().isoformat()
+            except ValueError:
+                continue
+            tokens.append((match.start(), match.end(), canonical))
+
+        for match in _SUMMARY_CHINESE_CALENDAR_DATE_RE.finditer(text):
+            try:
+                year = int(match.group("year")) if match.group("year") else (
+                    timestamp.year if timestamp is not None else None
+                )
+                if year is None:
+                    continue
+                canonical = datetime(
+                    year, int(match.group("month")), int(match.group("day")), tzinfo=timezone.utc
+                ).date().isoformat()
+            except ValueError:
+                continue
+            tokens.append((match.start(), match.end(), canonical))
+
+        for match in _SLASH_CALENDAR_DATE_RE.finditer(text):
+            try:
+                year = int(match.group("year")) if match.group("year") else (
+                    timestamp.year if timestamp is not None else None
+                )
+                if year is None:
+                    continue
+                canonical = datetime(
+                    year, int(match.group("month")), int(match.group("day")), tzinfo=timezone.utc
+                ).date().isoformat()
+            except ValueError:
+                continue
+            tokens.append((match.start(), match.end(), canonical))
+
+        for start, end, canonical in tokens:
+            prefix = text[max(0, start - 32):start]
+            suffix = text[end:end + 24]
+            if _DEADLINE_PREFIX_RE.search(prefix) or _DEADLINE_SUFFIX_RE.search(suffix):
+                result.add(canonical)
+    return result
+
+
 def _native_result(value: Any) -> list[dict[str, Any]]:
     if value is None:
         return []

{memleaf-0.2.53 → memleaf-0.2.55}/src/memleaf/single_pass_memory_planner.py

@@ -24,6 +24,8 @@ from .process_common import (
     ProcessingError,
     _automatic_create_conflicts,
     _event_payload,
+    _explicit_project_scope_labels,
+    _grounded_deadline_dates,
     _grounded_due_dates,
     _normalize_summary_dates,
     _summary_date_grounding_violations,
@@ -53,6 +55,87 @@ def _drop_ungrounded_project_scopes(scopes: Iterable[Any]) -> list[str]:
     return kept or ["global"]
 
 
+def _claim_date_evidence(
+    claims: Iterable[Mapping[str, Any]],
+    by_unit: Mapping[str, Any],
+    events: Iterable[Mapping[str, Any]],
+) -> list[dict[str, Any]]:
+    """Project dates from this candidate's canonical validated exact quotes only."""
+
+    event_by_key = {
+        event.get("event_key"): event
+        for event in events
+        if isinstance(event, Mapping) and isinstance(event.get("event_key"), str)
+    }
+    result: list[dict[str, Any]] = []
+    seen: set[tuple[str, str]] = set()
+    for claim in claims:
+        if not isinstance(claim, Mapping):
+            continue
+        unit_id = claim.get("unit_id")
+        quote = claim.get("quote")
+        unit = by_unit.get(unit_id) if isinstance(unit_id, str) else None
+        if not isinstance(quote, str) or not quote or unit is None:
+            continue
+        event = event_by_key.get(getattr(unit, "event_key", None))
+        if event is None or getattr(unit, "source_role", None) not in {"user", "assistant"}:
+            continue
+        identity = (unit_id, quote)
+        if identity in seen:
+            continue
+        seen.add(identity)
+        result.append({
+            "unit_id": unit_id,
+            "event_key": getattr(unit, "event_key", None),
+            "role": getattr(unit, "source_role", None),
+            "timestamp": event.get("timestamp"),
+            "content": quote,
+        })
+    return result
+
+
+def _fill_unambiguous_todo_deadline(
+    memory_type: Any,
+    summary: Mapping[str, Any],
+    deadline_dates: Iterable[str],
+) -> tuple[dict[str, Any], bool]:
+    """Fill an omitted todo date only when candidate evidence proves one deadline."""
+
+    normalized = dict(summary)
+    dates = {value for value in deadline_dates if isinstance(value, str)}
+    if memory_type != "todo" or normalized.get("due_date") is not None:
+        return normalized, False
+    if len(dates) == 1:
+        normalized["due_date"] = next(iter(dates))
+        return normalized, False
+    return normalized, len(dates) > 1
+
+
+def _global_scope_conflicts_with_candidate_evidence(
+    scopes: Iterable[Any],
+    candidate_evidence: Iterable[Mapping[str, Any]],
+    by_unit: Mapping[str, Any],
+    scope_registry: Mapping[str, Any] | None,
+) -> bool:
+    """Detect a global choice contradicted by one explicit project label."""
+
+    scope_values = {value.casefold() for value in scopes if isinstance(value, str)}
+    if "global" not in scope_values or any(value.startswith("project:") for value in scope_values):
+        return False
+    texts: list[str] = []
+    for event in candidate_evidence:
+        if not isinstance(event, Mapping):
+            continue
+        unit = by_unit.get(event.get("unit_id"))
+        section_path = getattr(unit, "section_path", ())
+        if isinstance(section_path, (list, tuple)):
+            texts.extend(value for value in section_path if isinstance(value, str))
+        content = event.get("content")
+        if isinstance(content, str):
+            texts.append(content)
+    return len(_explicit_project_scope_labels(texts, scope_registry)) == 1
+
+
 class SinglePassMemoryPlanner(MemoryPlanner):
     """One semantic model call for an ordinary automatic turn."""
 
@@ -350,6 +433,7 @@ class SinglePassMemoryPlanner(MemoryPlanner):
         no_memory_by_unit: dict[str, str] = {}
         observed_scopes: list[str] = []
         requests: list[dict[str, Any]] = []
+        ambiguous_todo_deadlines: set[str] = set()
 
         if not planning_units:
             self._finalize_evidence_audit(
@@ -443,6 +527,18 @@ class SinglePassMemoryPlanner(MemoryPlanner):
             }
             if decision == "UPDATE" and isinstance(target_id, str):
                 candidate["update_memory_id"] = target_id
+
+            candidate_evidence = _claim_date_evidence(claims, by_unit, events)
+            if _global_scope_conflicts_with_candidate_evidence(
+                scopes, candidate_evidence, by_unit, validation_scope_registry
+            ):
+                # The evidence explicitly labels one project, but Core cannot
+                # safely rewrite the model's chosen scope. Defer only this
+                # candidate rather than persist it globally.
+                raise ModelOutputError(
+                    "global scope conflicts with an explicit project label",
+                    validation_detail="scope_not_grounded",
+                )
             if decision in {"CREATE", "UPDATE"} and not _model_project_scope_is_source_grounded(
                 candidate,
                 planning_units,
@@ -500,7 +596,10 @@ class SinglePassMemoryPlanner(MemoryPlanner):
                 event["event_key"] for event in admitted_events
                 if isinstance(event, Mapping) and isinstance(event.get("event_key"), str)
             ))
-            grounded_dates = _grounded_due_dates(turn, evidence_events=admitted_events)
+            candidate_date_evidence = candidate_evidence
+            grounded_dates = _grounded_due_dates(turn, evidence_events=candidate_date_evidence)
+            deadline_dates = _grounded_deadline_dates(candidate_date_evidence)
+            grounded_dates.update(deadline_dates)
             summary = dict(proposed)
             if decision == "UPDATE" and target_memory is not None:
                 summary.setdefault("title", target_memory.title)
@@ -512,6 +611,11 @@ class SinglePassMemoryPlanner(MemoryPlanner):
                 summary["update_memory_id"] = target_memory.memory_id
             else:
                 summary.setdefault("tags", [])
+            summary, deadline_is_ambiguous = _fill_unambiguous_todo_deadline(
+                memory_type, summary, deadline_dates
+            )
+            if deadline_is_ambiguous:
+                ambiguous_todo_deadlines.add(candidate_id)
             summary["type"] = memory_type
             summary["scopes"] = scopes
             if isinstance(scope_source, str) and scope_source:
@@ -535,7 +639,7 @@ class SinglePassMemoryPlanner(MemoryPlanner):
                 expected_target_type=target_memory.type if target_memory is not None else None,
                 expected_scopes=scopes,
                 expected_scope_source=scope_source if isinstance(scope_source, str) else None,
-                allowed_due_dates=grounded_dates,
+                allowed_due_dates=(deadline_dates or grounded_dates),
                 allow_no_change=False,
                 allow_update_target=target_memory is not None,
             )
@@ -543,8 +647,7 @@ class SinglePassMemoryPlanner(MemoryPlanner):
                 parsed,
                 grounded_dates=grounded_dates,
                 source_texts=[
-                    event.get("content", "") for event in admitted_events
-                    if isinstance(event, Mapping) and event.get("role") in {"user", "assistant"}
+                    event.get("content", "") for event in candidate_date_evidence
                 ],
                 preserved_texts=(
                     target_memory.title,
@@ -581,6 +684,15 @@ class SinglePassMemoryPlanner(MemoryPlanner):
             },
         )
 
+        if ambiguous_todo_deadlines:
+            metric = getattr(self.model, "_record_metric_event", None)
+            if callable(metric):
+                for _ in ambiguous_todo_deadlines:
+                    metric(
+                        {"stage": "single_pass", "operation": "single_pass_primary"},
+                        "b3_due_date_ambiguous_count",
+                    )
+
         fallback_scopes, fallback_scope_source = self._fallback_scopes(scope_background)
         for row in result["items"]:
             candidate_id = row["candidate_id"]

{memleaf-0.2.53 → memleaf-0.2.55}/src/memleaf/single_pass_plan.py

@@ -9,6 +9,7 @@ from __future__ import annotations
 
 import json
 import re
+import unicodedata
 from copy import deepcopy
 from decimal import Decimal, InvalidOperation
 from collections.abc import Callable, Iterable, Mapping, Sequence
@@ -95,6 +96,24 @@ _DECISION_OPTIONAL_FIELDS = {
 _LEGACY_REDUNDANT_ITEM_FIELDS = frozenset({"sources", "update_memory_id"})
 _LEGACY_REDUNDANT_MEMORY_FIELDS = frozenset({"type", "scopes", "scope_source", "sources", "update_memory_id"})
 _B3_REPAIR_MAX_BYTES = 64 * 1024
+_TARGET_SENTENCE_SPLIT = re.compile(r"[\r\n。！？!?；;]+")
+_TARGET_ASCII_ANCHOR = re.compile(r"[a-z0-9]+(?:[._/-][a-z0-9]+)*")
+_TARGET_CJK_RUN = re.compile(r"[\u3400-\u4dbf\u4e00-\u9fff\uf900-\ufaff]+")
+_TARGET_DATE_ANCHOR = re.compile(
+    r"(?<![a-z0-9])(?:\d{4}[-/.年]\d{1,2}[-/.月]\d{1,2}日?)(?![a-z0-9])"
+)
+_TARGET_GENERIC_CJK_BIGRAMS = frozenset({
+    "用户", "我们", "他们", "自己", "要求", "需要", "应该", "可以", "如果", "因为",
+    "所以", "没有", "不是", "已经", "目前", "现在", "相关", "问题", "情况", "项目",
+    "一个", "一些", "这个", "那个", "内容", "事情", "工作", "处理", "本次", "通过",
+    "根据", "后续", "其他", "结果", "暂时", "异常", "注意", "确认", "完成", "说明",
+    "状态", "新增", "进行", "是否", "可能", "所有", "每个", "提供", "发生",
+})
+_TARGET_GENERIC_CJK_PHRASES = frozenset({
+    "用户要求", "后续需要", "需要确认", "目前没有", "没有问题", "没有异常", "无异常情况",
+    "项目相关", "相关项目", "进行处理", "重要内容", "邮件内容", "本次处理", "需要进一步",
+    "工作内容", "结果如下", "如何处理", "用户需要", "继续跟进", "确认状态", "是否需要",
+})
 
 
 def _enum_text(values: Iterable[str]) -> str:
@@ -109,7 +128,7 @@ UPDATE exactly requires candidate_id,decision,evidence,target_memory_id,memory;
 NO_CHANGE exactly requires candidate_id,decision,evidence,target_memory_id.
 DEFERRED exactly requires candidate_id,decision,evidence,reason; reason={_enum_text(_DEFER_REASONS)}.
 Memory allowed only: title,body,tags,aliases,keywords,status,completed_at,due_date,shadow_native_ids. status={_enum_text(TODO_STATUSES)}. status,completed_at,due_date are todo-only: omit all three for every other type. A todo UPDATE must restate its current status; status=completed requires completed_at and completed_at requires status=completed. Never put type,scopes,scope_source,sources,update_memory_id in memory.
-Evidence is a NONEMPTY ARRAY of claims, never one bare claim object. Each claim is exactly one of: {{unit_id,quote,role}} OR {{unit_id,whole_unit:true,role}} OR {{unit_id,start,end,quote,role}}. role={_enum_text(_EVIDENCE_ROLES)}. Offsets are start-inclusive/end-exclusive and text[start:end]==quote; quote-only must occur exactly once; user_confirmation must cite user evidence.
+Evidence is a NONEMPTY ARRAY of claims, never one bare claim object. Each claim is exactly one of: {{unit_id,quote,role}} OR {{unit_id,whole_unit:true,role}} OR {{unit_id,start,end,quote,role}}. role={_enum_text(_EVIDENCE_ROLES)}. Offsets are start-inclusive/end-exclusive and text[start:end]==quote; quote-only must occur exactly once; user_confirmation must cite user evidence. Prefer the shortest exact quote; for a long reply with several facts, avoid whole_unit when a specific span suffices.
 NoMemory row exactly {{unit_id,reason}}; reason={_enum_text(_NO_MEMORY_REASONS)}.
 Every current_evidence unit must be claimed by >=1 item OR appear exactly once in no_memory, never both and never omitted. One evidence unit may support multiple independent items.
 CREATE/UPDATE/NO_CHANGE require lookup_complete=true. UPDATE/NO_CHANGE target only local_memory_catalog; a target may be used once: when several changes touch one target, emit ONE UPDATE carrying their merged current state, never several items for the same target.
@@ -120,20 +139,17 @@ Return one JSON object only. No Markdown, explanation, or reasoning."""
 SINGLE_PASS_SYSTEM = f"""You are memleaf's single-pass memory planner.
 
 SOURCE
-Only current_evidence may establish new facts or changes. local_memory_catalog and native_memory_catalog are comparison context; native memory is never an UPDATE/NO_CHANGE target. Never invent source facts, ownership, dates, status, obligations, numbers, relationships or IDs.
-An assistant message is evidence of what you reported, never of what the user wants. It may organise, restate or report what the user's own evidence establishes, but it may never be the only support for a durable claim. When it states something the user's message does not establish -- your own decision, requirement, rule, plan, recommendation or added detail -- record nothing from that statement, however definite it sounds. Consequences you derived from a user preference are yours, not the user's: never record them as a requirement or as project state. Only something the user stated or accepted becomes memory.
+Use current_evidence only. User messages and assistant final reports may support facts, including facts obtained from external sources. Do not turn your own advice, plans or inferences into user intent unless accepted. Never invent facts, dates, numbers, ownership or IDs. local_memory_catalog is comparison context; native memory is never an UPDATE/NO_CHANGE target.
 
 TASK
-Extract every independently useful long-term memory. Keep independently retrievable/updateable topics separate and preserve entity, condition, polarity, uncertainty, ownership, state and meaning-critical numbers/codes. CREATE only when no supplied local memory represents the durable information; UPDATE only when current evidence proves a change to one supplied local memory; NO_CHANGE only when it adds no semantic change; DEFERRED for a durable candidate that cannot safely reach a terminal decision. Project ownership and platform/system names are separate judgments. Do not turn every negation into no_memory and do not use NO_CHANGE to hide ambiguity.
+Keep independently retrievable facts useful for future answers, actions, commitments, status tracking or avoiding repeated research. Prefer stable facts, decisions, open work and deadlines; skip transient failures, one-time fallbacks, routine checks with no follow-up, and point-in-time counts or snapshots unless needed for a trend, threshold, obligation, decision or later comparison. Preserve future-use facts in final reports and keep independent topics separate. Preserve entity, condition, polarity, uncertainty, ownership, state and meaning-critical numbers/codes. Use self-contained wording; call the conversation person “the user” (用户), never “owner” (主人). CREATE only if no local memory represents the information; UPDATE only for a proven change to one target; NO_CHANGE only for the same future-use item with no semantic change; DEFERRED for an unsafe terminal decision. Do not use NO_CHANGE to hide ambiguity.
 
 SCOPES
 Legal values: global | domain:<name> | portfolio:<name> | project:<name> | unscoped. scopes is a nonempty array; at most one project:<name> per memory; unscoped must be the only value, and Core then records insufficient_context.
-When a candidate concerns one distinct project the user is working on -- a type=project memory, or the work, decisions, state or deadlines that belong to that project -- scope it to that project as project:<name>, using the name the evidence itself uses. The project does not need to be registered first; scope_registry only lists what already exists.
-Ground the name in this candidate's own cited evidence, or in a project scope explicitly supplied for this turn. Never edit, translate or combine the evidence's words into a name that does not appear there, and choose the shortest term that does appear and names the project: the thing described as "记账的小玩意儿" is project:记账, not project:记账小玩意儿 and not project:记账的小玩意儿. An ungrounded name is rejected and its memory is deferred instead of written. A product, platform, system, vendor, notification source, comparison or implementation context is not ownership by name alone.
-Use global only for a fact that no single project owns: a standing personal preference, a machine-wide or tool-wide rule, or an environment fact. Never invent, translate or borrow a project name.
+Scope to one project only when this candidate's evidence identifies it as owner of the work, decision, state or deadline; use a name present in that evidence, whether or not registered. A platform, product or vendor mention alone is not ownership; defer when ownership is unclear. Use global for cross-project facts, standing preferences and environment-wide rules.
 
 DATES
-An evidence unit may carry an ISO-8601 UTC timestamp. Use it ONLY to resolve a relative, partial or yearless date that the unit's own text expresses; never borrow another unit's timestamp and never guess a missing year. The timestamp is an anchor, not content: never write its own date into a memory, and add no date the evidence text does not state. A date literal in a memory must appear in that memory's cited evidence, either verbatim or as the same month and day. Write due_date as YYYY-MM-DD. A todo declares status, and due_date only when its own evidence establishes a deadline: otherwise omit due_date, or set it to null when the deadline itself is still unresolved. Any memory carrying a date that no admitted evidence grounds is rejected and costs the whole turn, so defer instead of approximating.
+An evidence unit may carry an ISO-8601 UTC timestamp. Use it ONLY to resolve a relative, partial or yearless date that the unit's own text expresses; never borrow another unit's timestamp and never guess a missing year. The timestamp is an anchor, not content: never write its own date into a memory, and add no date the evidence text does not state. A date literal in a memory must appear in that memory's cited evidence, either verbatim or as the same month and day. Write due_date as YYYY-MM-DD; a todo with one unambiguous evidence-grounded deadline must include it. Omit due_date when no deadline is stated or it cannot be resolved. Any memory carrying a date that no admitted evidence grounds is rejected and costs the whole turn, so defer instead of approximating.
 
 {B3_COMPACT_CONTRACT}"""
 
@@ -420,6 +436,71 @@ def _canonical_target(raw: Any, local_by_key: Mapping[str, Mapping[str, Any]]) -
     return target["memory_id"], target
 
 
+def _target_anchor_sets(text: str) -> tuple[set[str], set[str]]:
+    normalized = unicodedata.normalize("NFKC", text).casefold()
+    without_dates = _TARGET_DATE_ANCHOR.sub(" ", normalized)
+    ascii_anchors = {
+        token for token in _TARGET_ASCII_ANCHOR.findall(without_dates)
+        if len(token) >= 3 or (token.isdigit() and len(token) >= 2)
+    }
+    cjk_anchors: set[str] = set()
+    for run in _TARGET_CJK_RUN.findall(normalized):
+        cjk_anchors.update(
+            run[index:index + 2]
+            for index in range(len(run) - 1)
+            if run[index:index + 2] not in _TARGET_GENERIC_CJK_BIGRAMS
+        )
+    return ascii_anchors, cjk_anchors
+
+
+def _target_is_same_future_use(
+    target: Mapping[str, Any], claims: Iterable[Mapping[str, Any]],
+) -> bool:
+    """Check a NO_CHANGE target against this candidate's exact cited text."""
+
+    target_units = [
+        part.strip()
+        for value in (target.get("title"), target.get("body"))
+        if isinstance(value, str)
+        for part in _TARGET_SENTENCE_SPLIT.split(value)
+        if part.strip()
+    ]
+    for claim in claims:
+        quote = claim.get("quote")
+        if not isinstance(quote, str) or not quote.strip():
+            continue
+        for quote_unit in (part.strip() for part in _TARGET_SENTENCE_SPLIT.split(quote) if part.strip()):
+            quote_ascii, quote_cjk = _target_anchor_sets(quote_unit)
+            for target_unit in target_units:
+                target_ascii, target_cjk = _target_anchor_sets(target_unit)
+                shared_ascii = quote_ascii & target_ascii
+                shared_cjk = quote_cjk & target_cjk
+                shared_numbers = {token for token in shared_ascii if token.isdigit()}
+                shared_identifiers = {token for token in shared_ascii if not token.isdigit()}
+                strong_identifiers = {
+                    token for token in shared_identifiers
+                    if len(token) >= 6 or any(char.isdigit() for char in token)
+                }
+                if len(strong_identifiers) >= 2 or (
+                    strong_identifiers and (shared_cjk or len(shared_identifiers) >= 2)
+                ):
+                    return True
+                if any(len(token) >= 5 for token in shared_numbers) and (
+                    shared_cjk or shared_identifiers
+                ):
+                    return True
+                if len(shared_cjk) < 2:
+                    continue
+                overlap = len(shared_cjk) / min(len(quote_cjk), len(target_cjk))
+                compact = "".join(
+                    char for char in unicodedata.normalize("NFKC", quote_unit).casefold()
+                    if char.isalnum()
+                )
+                if overlap >= 0.28 and compact not in _TARGET_GENERIC_CJK_PHRASES:
+                    return True
+    return False
+
+
 MemoryValidator = Callable[
     [str, str, str | None, Mapping[str, Any] | None, Mapping[str, Any], list[dict[str, Any]], Mapping[str, Any]],
     Mapping[str, Any],
@@ -498,6 +579,8 @@ def parse_single_pass_output(
     binding_rows: list[dict[str, Any]] = []
     prepared: list[tuple[dict[str, Any], Mapping[str, Any] | None]] = []
     forced_defer: dict[str, str] = {}
+    broad_whole_unit_no_change: set[str] = set()
+    unit_by_id = {getattr(unit, "unit_id", None): unit for unit in source_units}
 
     for item_index, raw_item in enumerate(items):
         item_path = f"items[{item_index}]"
@@ -579,6 +662,16 @@ def parse_single_pass_output(
                 path=f"{item_path}.evidence", rule="type", actual=claims,
                 expected_type="array",
             )
+        if decision == "NO_CHANGE" and any(
+            isinstance(claim, Mapping)
+            and claim.get("whole_unit") is True
+            and len(getattr(unit_by_id.get(claim.get("unit_id")), "text", "")) > 512
+            for claim in claims
+        ):
+            # A whole long reply can contain many independent topics. It is
+            # insufficient proof that one selected old memory covers this
+            # candidate; ask for a narrow claim on a later bounded pass.
+            broad_whole_unit_no_change.add(candidate_key)
         binding_rows.append({"candidate_id": candidate_id, "claims": claims})
 
         target_record: Mapping[str, Any] | None = None
@@ -790,7 +883,27 @@ def parse_single_pass_output(
                 normalized["target_memory_id"] = target_id
                 normalized["memory"] = dict(validated)
             elif decision == "NO_CHANGE":
-                normalized["target_memory_id"] = item["target_memory_id"]
+                if (
+                    candidate_id.casefold() in broad_whole_unit_no_change
+                    or not isinstance(target_record, Mapping)
+                    or not _target_is_same_future_use(
+                        target_record, evidence
+                    )
+                ):
+                    normalized = {
+                        "candidate_id": candidate_id,
+                        "decision": "DEFERRED",
+                        "reason": "target_ambiguous",
+                        "evidence": evidence,
+                    }
+                    if deferrals is not None:
+                        deferrals.append({
+                            "candidate_id": candidate_id,
+                            "reason": "target_ambiguous",
+                            "detail": "target_relevance_unproven",
+                        })
+                else:
+                    normalized["target_memory_id"] = item["target_memory_id"]
             else:
                 normalized["reason"] = item["reason"]
         except ModelOutputError as error:

{memleaf-0.2.53 → memleaf-0.2.55}/src/memleaf/validation.py

@@ -344,7 +344,7 @@ _RELATIVE_DATE_TOKEN = (
     r"(?<![A-Za-z])(?:today|tomorrow|yesterday)(?![A-Za-z])"
     r"|(?<![A-Za-z])(?:this|next|last)\s+(?:monday|tuesday|wednesday|thursday|friday|saturday|sunday)(?![A-Za-z])"
     r"|(?:本|这|下|上)(?:个)?(?:周|星期|礼拜)\s*"
-    r"(?:(?:星期|礼拜)\s*)?(?:一|二|三|四|五|六|日|天|末|[1-7])"
+    r"(?:(?:星期|礼拜)\s*)?(?:一|二|三|四|五|六|天|末|[1-7]|日(?!报))"
     r"|(?:今天|明天|昨天|今日|明日|昨日)"
     r")"
 )
@@ -364,6 +364,9 @@ _ISO_CALENDAR_DATE = re.compile(
     r"(?<![A-Za-z\d./-])\d{4}-(?:0?[1-9]|1[0-2])-(?:0?[1-9]|[12]\d|3[01])"
     r"(?![A-Za-z\d./-])"
 )
+_ISO_DATE_FOLLOWED_BY_CLOCK = re.compile(
+    r"(?P<date>\d{4}-\d{2}-\d{2})(?=(?:[01]?\d|2[0-3]):[0-5]\d(?::[0-5]\d)?(?!\d))"
+)
 _EMPTY_ISO_DATE_PARENTHESIS = re.compile(
     r"(?P<date>\d{4}-(?:0?[1-9]|1[0-2])-(?:0?[1-9]|[12]\d|3[01]))"
     r"[ \t]*(?:\([ \t]*[)）]|（[ \t]*[)）])"
@@ -605,6 +608,18 @@ def _normalize_relative_calendar_text(
     normalized = _RELATIVE_CALENDAR_EXPRESSION.sub(replace_token, text)
     if relative_replaced:
         normalized = _collapse_duplicate_calendar_dates(normalized)
+    # A relative date can be adjacent to a clock in compact source text (for
+    # example, "today10:30"). Keep the date and time as separate values after
+    # normalization, while leaving already-spaced dates unchanged.
+    def separate_date_from_clock(match: re.Match[str]) -> str:
+        value = match.group("date")
+        try:
+            datetime.strptime(value, "%Y-%m-%d")
+        except ValueError:
+            return match.group(0)
+        return value + " "
+
+    normalized = _ISO_DATE_FOLLOWED_BY_CLOCK.sub(separate_date_from_clock, normalized)
     normalized = _strip_empty_iso_date_parenthesis(normalized)
     return normalized, safe
 

{memleaf-0.2.53 → memleaf-0.2.55/src/memleaf.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: memleaf
-Version: 0.2.53
+Version: 0.2.55
 Summary: A local-first Markdown memory core for AI agents
 Author: memleaf contributors
 License-Expression: MIT
@@ -23,7 +23,7 @@ Dynamic: license-file
 
 [English](README.en.md) · [PyPI](https://pypi.org/project/memleaf/) · [GitHub](https://github.com/miffyblueboo/memleaf)
 
-> **版本：0.2.53。**
+> **版本：0.2.55。**
 > 记忆只从用户与 Agent 的可见对话提炼。Agent 已在回复中整理的事实、项目进展和明确待办可作为来源；邮件、附件、网页和终端等工具原文不进入记忆提炼。模型负责保留原话中的不确定性，现有记忆仅用于比较、去重和更新。Markdown 仍是唯一事实源。
 > **当前版本支持 Hermes 和 Codex。** Antigravity（反重力）不检测、不安装、不配置。