biblicus 0.16.0__py3-none-any.whl → 1.0.0__py3-none-any.whl

biblicus/corpus.py

@@ -11,6 +11,7 @@ import shutil
 import uuid
 from pathlib import Path
 from typing import Any, Dict, List, Optional, Sequence
+from urllib.parse import quote, unquote, urlparse
 
 import yaml
 from pydantic import ValidationError
@@ -24,6 +25,7 @@ from .constants import (
     SCHEMA_VERSION,
     SIDECAR_SUFFIX,
 )
+from .errors import IngestCollisionError
 from .frontmatter import parse_front_matter, render_front_matter
 from .hook_manager import HookManager
 from .hooks import HookPoint
@@ -110,7 +112,10 @@ def _preferred_extension_for_media_type(media_type: str) -> Optional[str]:
     """
     media_type_overrides = {
         "image/jpeg": ".jpg",
+        "audio/mpeg": ".mp3",
         "audio/ogg": ".ogg",
+        "audio/wav": ".wav",
+        "audio/x-wav": ".wav",
     }
     if media_type in media_type_overrides:
         return media_type_overrides[media_type]
@@ -136,7 +141,16 @@ def _ensure_filename_extension(filename: str, *, media_type: str) -> str:
         return raw_name + ".md"
 
     if Path(raw_name).suffix:
-        return raw_name
+        if "%2F" in raw_name or "%3A" in raw_name:
+            decoded = unquote(raw_name)
+            parsed = urlparse(decoded)
+            decoded_path = parsed.path if parsed.scheme else decoded
+            if not Path(decoded_path).suffix:
+                pass
+            else:
+                return raw_name
+        else:
+            return raw_name
 
     ext = _preferred_extension_for_media_type(media_type)
     if not ext:
@@ -144,6 +158,55 @@ def _ensure_filename_extension(filename: str, *, media_type: str) -> str:
     return raw_name + ext
 
 
+def _encode_source_uri_for_filename(source_uri: str) -> str:
+    """
+    Percent-encode a source uniform resource identifier for filename use.
+
+    :param source_uri: Source uniform resource identifier to encode.
+    :type source_uri: str
+    :return: Percent-encoded uniform resource identifier safe for filenames.
+    :rtype: str
+    """
+    return quote(source_uri, safe="")
+
+
+def _storage_filename_for_ingest(
+    *, filename: Optional[str], media_type: str, source_uri: Optional[str]
+) -> str:
+    """
+    Derive a collision-safe filename for corpus storage.
+
+    If a source uniform resource identifier is provided, the full uniform resource identifier is
+    percent-encoded to namespace the stored file, preventing collisions between identical basenames
+    from different sources. When no uniform resource identifier is available, fall back to a
+    sanitized filename.
+
+    :param filename: Optional filename hint from the caller.
+    :type filename: str or None
+    :param media_type: Media type of the payload.
+    :type media_type: str
+    :param source_uri: Optional source uniform resource identifier for provenance.
+    :type source_uri: str or None
+    :return: Storage filename with an appropriate extension, or an empty string when no hint exists.
+    :rtype: str
+    """
+    base_name = ""
+    if source_uri:
+        base_name = _encode_source_uri_for_filename(source_uri)
+        if filename and not source_uri.startswith("file:"):
+            sanitized = _sanitize_filename(filename)
+            if sanitized:
+                base_name = f"{base_name}--{sanitized}"
+    if not base_name and filename:
+        base_name = _sanitize_filename(filename)
+    if not base_name:
+        return ""
+    if len(base_name) > 180:
+        digest = hashlib.sha256(base_name.encode("utf-8")).hexdigest()
+        base_name = f"hash-{digest}"
+    return _ensure_filename_extension(base_name, media_type=media_type)
+
+
 def _merge_tags(explicit: Sequence[str], from_frontmatter: Any) -> List[str]:
     """
     Merge tags from explicit input and front matter values.
@@ -520,6 +583,24 @@ class Corpus:
         temp_path.write_text(catalog.model_dump_json(indent=2) + "\n", encoding="utf-8")
         temp_path.replace(self.catalog_path)
 
+    def _find_item_by_source_uri(self, source_uri: str) -> Optional[CatalogItem]:
+        """
+        Locate an existing catalog item by source uniform resource identifier.
+
+        :param source_uri: Source uniform resource identifier to search for.
+        :type source_uri: str
+        :return: Matching catalog item or None.
+        :rtype: CatalogItem or None
+        """
+        if not source_uri:
+            return None
+        self._init_catalog()
+        catalog = self._load_catalog()
+        for item in catalog.items.values():
+            if item.source_uri == source_uri:
+                return item
+        return None
+
     @property
     def runs_dir(self) -> Path:
         """
@@ -817,18 +898,26 @@ class Corpus:
         :return: Ingestion result summary.
         :rtype: IngestResult
         :raises ValueError: If markdown is not Unicode Transformation Format 8.
+        :raises IngestCollisionError: If a source uniform resource identifier is already ingested.
         """
-        item_id = str(uuid.uuid4())
-        safe_filename = _sanitize_filename(filename) if filename else ""
+        existing_item = self._find_item_by_source_uri(source_uri)
+        if existing_item is not None:
+            raise IngestCollisionError(
+                source_uri=source_uri,
+                existing_item_id=existing_item.id,
+                existing_relpath=existing_item.relpath,
+            )
 
-        if safe_filename:
-            safe_filename = _ensure_filename_extension(safe_filename, media_type=media_type)
+        item_id = str(uuid.uuid4())
+        storage_filename = _storage_filename_for_ingest(
+            filename=filename, media_type=media_type, source_uri=source_uri
+        )
 
         if media_type == "text/markdown":
-            output_name = f"{item_id}--{safe_filename}" if safe_filename else f"{item_id}.md"
+            output_name = f"{item_id}--{storage_filename}" if storage_filename else f"{item_id}.md"
         else:
-            if safe_filename:
-                output_name = f"{item_id}--{safe_filename}"
+            if storage_filename:
+                output_name = f"{item_id}--{storage_filename}"
             else:
                 extension = _preferred_extension_for_media_type(media_type) or ""
                 output_name = f"{item_id}{extension}" if extension else f"{item_id}"
@@ -991,13 +1080,21 @@ class Corpus:
         if media_type == "text/markdown":
             raise ValueError("Stream ingestion is not supported for Markdown")
 
+        existing_item = self._find_item_by_source_uri(source_uri)
+        if existing_item is not None:
+            raise IngestCollisionError(
+                source_uri=source_uri,
+                existing_item_id=existing_item.id,
+                existing_relpath=existing_item.relpath,
+            )
+
         item_id = str(uuid.uuid4())
-        safe_filename = _sanitize_filename(filename) if filename else ""
-        if safe_filename:
-            safe_filename = _ensure_filename_extension(safe_filename, media_type=media_type)
+        storage_filename = _storage_filename_for_ingest(
+            filename=filename, media_type=media_type, source_uri=source_uri
+        )
 
-        if safe_filename:
-            output_name = f"{item_id}--{safe_filename}"
+        if storage_filename:
+            output_name = f"{item_id}--{storage_filename}"
         else:
             extension = _preferred_extension_for_media_type(media_type) or ""
             output_name = f"{item_id}{extension}" if extension else f"{item_id}"
@@ -1085,7 +1182,7 @@ class Corpus:
         *,
         title: Optional[str] = None,
         tags: Sequence[str] = (),
-        source_uri: str = "text",
+        source_uri: Optional[str] = None,
     ) -> IngestResult:
         """
         Ingest a text note as Markdown.
@@ -1096,11 +1193,15 @@ class Corpus:
         :type title: str or None
         :param tags: Tags to associate with the note.
         :type tags: Sequence[str]
-        :param source_uri: Source uniform resource identifier for provenance.
-        :type source_uri: str
+        :param source_uri: Optional source uniform resource identifier for provenance.
+        :type source_uri: str or None
         :return: Ingestion result summary.
         :rtype: IngestResult
         """
+        if source_uri is None:
+            digest_source = (title or "") + "\n" + text
+            digest = hashlib.sha256(digest_source.encode("utf-8")).hexdigest()
+            source_uri = f"text:{digest}"
         data = text.encode("utf-8")
         return self.ingest_item(
             data,

biblicus/errors.py

@@ -13,3 +13,27 @@ class ExtractionRunFatalError(RuntimeError):
     rather than a per-item extraction failure. For example, a selection extractor that depends
     on referenced extraction run manifests treats missing manifests as fatal.
     """
+
+
+class IngestCollisionError(RuntimeError):
+    """
+    Ingest collision for an already ingested source.
+
+    :param source_uri: Source uniform resource identifier that caused the collision.
+    :type source_uri: str
+    :param existing_item_id: Identifier of the existing catalog item.
+    :type existing_item_id: str
+    :param existing_relpath: Raw storage relpath of the existing item.
+    :type existing_relpath: str
+    """
+
+    def __init__(self, *, source_uri: str, existing_item_id: str, existing_relpath: str) -> None:
+        self.source_uri = source_uri
+        self.existing_item_id = existing_item_id
+        self.existing_relpath = existing_relpath
+        message = (
+            "Source already ingested"
+            f": source_uri={source_uri} existing_item_id={existing_item_id}"
+            f" existing_relpath={existing_relpath}"
+        )
+        super().__init__(message)

biblicus/knowledge_base.py

@@ -44,7 +44,7 @@ class KnowledgeBaseDefaults(BaseModel):
     query_budget: QueryBudget = Field(
         default_factory=lambda: QueryBudget(
             max_total_items=5,
-            max_total_characters=2000,
+            maximum_total_characters=2000,
             max_items_per_source=None,
         )
     )

biblicus/models.py

@@ -234,8 +234,8 @@ class QueryBudget(BaseModel):
         This enables simple pagination by re-running the same query with a
         higher offset.
     :vartype offset: int
-    :ivar max_total_characters: Optional maximum total characters across evidence text.
-    :vartype max_total_characters: int or None
+    :ivar maximum_total_characters: Optional maximum total characters across evidence text.
+    :vartype maximum_total_characters: int or None
     :ivar max_items_per_source: Optional cap per source uniform resource identifier.
     :vartype max_items_per_source: int or None
     """
@@ -244,7 +244,7 @@ class QueryBudget(BaseModel):
 
     max_total_items: int = Field(ge=1)
     offset: int = Field(default=0, ge=0)
-    max_total_characters: Optional[int] = Field(default=None, ge=1)
+    maximum_total_characters: Optional[int] = Field(default=None, ge=1)
     max_items_per_source: Optional[int] = Field(default=None, ge=1)
 
 
@@ -278,6 +278,8 @@ class Evidence(BaseModel):
     :vartype recipe_id: str
     :ivar run_id: Retrieval run identifier.
     :vartype run_id: str
+    :ivar metadata: Optional metadata payload from the catalog item.
+    :vartype metadata: dict[str, Any]
     :ivar hash: Optional content hash for provenance.
     :vartype hash: str or None
     """
@@ -297,6 +299,7 @@ class Evidence(BaseModel):
     stage_scores: Optional[Dict[str, float]] = None
     recipe_id: str
     run_id: str
+    metadata: Dict[str, Any] = Field(default_factory=dict)
     hash: Optional[str] = None
 
     @model_validator(mode="after")

biblicus/retrieval.py

@@ -124,8 +124,8 @@ def apply_budget(evidence: Iterable[Evidence], budget: QueryBudget) -> List[Evid
                 continue
 
         text_character_count = len(candidate_evidence.text or "")
-        if budget.max_total_characters is not None:
-            if total_characters + text_character_count > budget.max_total_characters:
+        if budget.maximum_total_characters is not None:
+            if total_characters + text_character_count > budget.maximum_total_characters:
                 continue
 
         selected_evidence.append(candidate_evidence)

biblicus/sources.py

@@ -8,7 +8,7 @@ import mimetypes
 from dataclasses import dataclass
 from pathlib import Path
 from typing import Optional
-from urllib.parse import unquote, urlparse
+from urllib.parse import quote, unquote, urlparse
 from urllib.request import Request, urlopen
 
 
@@ -37,6 +37,27 @@ def _filename_from_url_path(path: str) -> str:
     return filename or "download"
 
 
+def _sanitize_filename_component(name: str) -> str:
+    allowed_characters = set("abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789-._() ")
+    sanitized_name = "".join(
+        (character if character in allowed_characters else "_") for character in name
+    ).strip()
+    return sanitized_name or "file"
+
+
+def _namespaced_filename(
+    *, source_uri: Optional[str], fallback_name: Optional[str], media_type: str
+) -> str:
+    base_name = ""
+    if source_uri:
+        base_name = quote(source_uri, safe="")
+    if not base_name and fallback_name:
+        base_name = _sanitize_filename_component(fallback_name)
+    if not base_name:
+        base_name = "file"
+    return _ensure_extension_for_media_type(base_name, media_type)
+
+
 def _media_type_from_filename(name: str) -> str:
     """
     Guess media type from a filename.
@@ -119,8 +140,16 @@ def _ensure_extension_for_media_type(filename: str, media_type: str) -> str:
     """
     if Path(filename).suffix:
         return filename
-    if media_type == "audio/ogg":
-        ext = ".ogg"
+    media_type_overrides = {
+        "audio/mpeg": ".mp3",
+        "audio/ogg": ".ogg",
+        "audio/wav": ".wav",
+        "audio/x-wav": ".wav",
+        "image/jpeg": ".jpg",
+        "text/html": ".html",
+    }
+    if media_type in media_type_overrides:
+        ext = media_type_overrides[media_type]
     else:
         ext = mimetypes.guess_extension(media_type) or ""
     return filename + ext if ext else filename
@@ -165,11 +194,12 @@ def load_source(source: str | Path, *, source_uri: Optional[str] = None) -> Sour
         media_type = _media_type_from_filename(path.name)
         if path.suffix.lower() in {".md", ".markdown"}:
             media_type = "text/markdown"
+        resolved_source_uri = source_uri or path.as_uri()
         return SourcePayload(
             data=path.read_bytes(),
             filename=path.name,
             media_type=media_type,
-            source_uri=source_uri or path.as_uri(),
+            source_uri=resolved_source_uri,
         )
 
     if _looks_like_uri(source):
@@ -187,21 +217,26 @@ def load_source(source: str | Path, *, source_uri: Optional[str] = None) -> Sour
             with urlopen(request, timeout=30) as response:
                 response_bytes = response.read()
                 content_type = response.headers.get("Content-Type", "").split(";", 1)[0].strip()
-                filename = _filename_from_url_path(parsed.path)
-                media_type = content_type or _media_type_from_filename(filename)
+                fallback_filename = _filename_from_url_path(parsed.path)
+                media_type = content_type or _media_type_from_filename(fallback_filename)
                 if media_type == "application/octet-stream":
                     sniffed = _sniff_media_type_from_bytes(response_bytes)
                     if sniffed:
                         media_type = sniffed
-                        filename = _ensure_extension_for_media_type(filename, media_type)
-                media_type = _normalize_media_type(filename=filename, media_type=media_type)
-                if Path(filename).suffix.lower() in {".md", ".markdown"}:
+                        fallback_filename = _ensure_extension_for_media_type(
+                            fallback_filename, media_type
+                        )
+                media_type = _normalize_media_type(
+                    filename=fallback_filename, media_type=media_type
+                )
+                if Path(fallback_filename).suffix.lower() in {".md", ".markdown"}:
                     media_type = "text/markdown"
+                resolved_source_uri = source_uri or source
                 return SourcePayload(
                     data=response_bytes,
-                    filename=filename,
+                    filename=fallback_filename,
                     media_type=media_type,
-                    source_uri=source_uri or source,
+                    source_uri=resolved_source_uri,
                 )
 
         raise NotImplementedError(

biblicus/text/link.py

@@ -159,6 +159,8 @@ def _apply_link_replace(text: str, old_str: str, new_str: str) -> str:
 
 
 def _validate_replace_text(old_str: str, new_str: str) -> None:
+    if "<span" in old_str or "</span>" in old_str:
+        raise ValueError("Text link replacements must target plain text without span tags")
     if strip_span_tags(old_str) != strip_span_tags(new_str):
         raise ValueError("Text link replacements may only insert span tags")
 
@@ -460,12 +462,16 @@ def _build_retry_message(errors: Sequence[str], current_text: str, id_prefix: st
     error_lines = "\n".join(f"- {error}" for error in errors)
     context_section = build_span_context_section(current_text, errors)
     coverage_guidance = _build_coverage_guidance(errors)
+    nested_guidance = ""
+    if any("nested span" in error for error in errors):
+        nested_guidance = "Do not create nested or overlapping spans. Remove nested spans and wrap only bare text.\n"
     return (
         "Your last edit did not validate.\n"
         "Issues:\n"
         f"{error_lines}\n\n"
         f"{context_section}"
         f"{coverage_guidance}"
+        f"{nested_guidance}"
         "Please fix the markup using str_replace. Use id for first mentions and ref for repeats. "
         "Reuse the same id for identical names and do not assign multiple ids to the same name. "
         f"Ids must start with '{id_prefix}'. Try again.\n"

biblicus/text/prompts.py

@@ -57,6 +57,8 @@ DEFAULT_ANNOTATE_SYSTEM_PROMPT = (
     "- new_str must be identical to old_str with only <span ...> and </span> inserted.\n"
     "- Do not include <span or </span> inside old_str or new_str.\n"
     "- Do not insert nested spans.\n"
+    "- Do not wrap text that is already inside a span; spans must never overlap.\n"
+    "- If a name appears inside an existing span, leave it alone and wrap only bare text.\n"
     "- If a tool call fails due to non-unique old_str, retry with a longer unique old_str.\n"
     "- If a tool call fails, read the error and keep editing. Do not call done until spans are inserted.\n"
     "- Do not delete, reorder, paraphrase, or label text beyond the span attributes.\n\n"

{biblicus-0.16.0.dist-info → biblicus-1.0.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: biblicus
-Version: 0.16.0
+Version: 1.0.0
 Summary: Command line interface and Python library for corpus ingestion, retrieval, and evaluation.
 License: MIT
 Requires-Python: >=3.9
@@ -293,7 +293,7 @@ for note_title, note_text in notes:
 
 backend = get_backend("scan")
 run = backend.build_run(corpus, recipe_name="Story demo", config={})
-budget = QueryBudget(max_total_items=5, max_total_characters=2000, max_items_per_source=None)
+budget = QueryBudget(max_total_items=5, maximum_total_characters=2000, max_items_per_source=None)
 result = backend.query(
     corpus,
     run=run,
@@ -333,7 +333,7 @@ Example output:
   "query_text": "Primary button style preference",
   "budget": {
     "max_total_items": 5,
-    "max_total_characters": 2000,
+    "maximum_total_characters": 2000,
     "max_items_per_source": null
   },
   "run_id": "RUN_ID",

{biblicus-0.16.0.dist-info → biblicus-1.0.0.dist-info}/RECORD

@@ -1,13 +1,13 @@
-biblicus/__init__.py,sha256=VK1nvxxf1NI1u5Ad94yyMpC5Xc_HJ7-3F_TnJdhItbA,496
+biblicus/__init__.py,sha256=z9Wif5-ZzIrptsUS8OELW5zG5_R3-4ZcSuVUkfqKbaA,989
 biblicus/__main__.py,sha256=ipfkUoTlocVnrQDM69C7TeBqQxmHVeiWMRaT3G9rtnk,117
 biblicus/chunking.py,sha256=GdJr0skAAI0Su99mr7dXqCgR7eJ0sJu8n2XesVGyddY,13206
-biblicus/cli.py,sha256=x3bbtg_nzvIZlHmiPp-4L2EtV6wugTMueFTkXQy9y1s,43372
+biblicus/cli.py,sha256=DdEL8Uvl38Zn2w4egCxQ4zWNelrI3QDs4qh4tGWGuAI,43793
 biblicus/constants.py,sha256=gAlEVJhxdFj-eWWJrlYbP7H1X3c5gwhrIBq9NQ1Vq_E,371
-biblicus/context.py,sha256=iXRFGpf_5YDPsDsm_iTK6nCvtUWDoYVI7op-l2QU3uA,10189
-biblicus/corpus.py,sha256=qSDnYJXhWlF2p_BbFLl6xtI53lIIPxwyKLLGLC432Sg,55612
+biblicus/context.py,sha256=I7L86ag2AbNr_QgiP5YSt1uwwULGx1cH73eR2nE9T3g,10842
+biblicus/corpus.py,sha256=LySjqBpTF_B19nMyGBoeB8AMDlqohcgsBfmJILm3P5c,59546
 biblicus/crawl.py,sha256=n8rXBMnziBK9vtKQQCXYOpBzqsPCswj2PzVJUb370KY,6250
 biblicus/embedding_providers.py,sha256=phWEsq1vryyTFRRs6uZ0sx9FhrqWIkDsS3I52I64zqM,3839
-biblicus/errors.py,sha256=uMajd5DvgnJ_-jq5sbeom1GV8DPUc-kojBaECFi6CsY,467
+biblicus/errors.py,sha256=g5TRPdO2XGi-7Wi1C4CXMJ6dTQKYAyP--EWKCv6FGKs,1362
 biblicus/evaluation.py,sha256=5xWpb-8f49Osh9aHzo1ab3AXOmls3Imc5rdnEC0pN-8,8143
 biblicus/evidence_processing.py,sha256=sJe6T1nLxvU0xs9yMH8JZZS19zHXMR-Fpr5lWi5ndUM,6120
 biblicus/extraction.py,sha256=qvrsq6zSz2Kg-cap-18HPHC9pQlqEGo7pyID2uKCyBo,19760
@@ -18,11 +18,11 @@ biblicus/hook_manager.py,sha256=ZCAkE5wLvn4lnQz8jho_o0HGEC9KdQd9qitkAEUQRcw,6997
 biblicus/hooks.py,sha256=OHQOmOi7rUcQqYWVeod4oPe8nVLepD7F_SlN7O_-BsE,7863
 biblicus/ignore.py,sha256=fyjt34E6tWNNrm1FseOhgH2MgryyVBQVzxhKL5s4aio,1800
 biblicus/inference.py,sha256=_k00AIPoXD2lruiTB-JUagtY4f_WKcdzA3axwiq1tck,3512
-biblicus/knowledge_base.py,sha256=JmlJw8WD_fgstuq1PyWVzU9kzvVzyv7_xOvhS70xwUw,6654
-biblicus/models.py,sha256=5AQ6oXK_KJyU0Kyv5ff8yD8nevNKb_6Hjr2_vlRSlK0,16297
+biblicus/knowledge_base.py,sha256=jpFEvo8gbEuwRUVYRRgQFvRTJZQml0WCHWSeY-CS4ag,6658
+biblicus/models.py,sha256=nvuq5Y96hHvuhMCuHff38wNITyQJam6zFrgFxH5Kh7g,16475
 biblicus/recipes.py,sha256=rqU66QnjOup6O8Y9Yq7XszmpoM0Pyrjw3RrfdnlVqgE,4210
-biblicus/retrieval.py,sha256=GXYT_3RPdqZEYdBQ4F4lIXDOhWw0nfL9bd781bgrn_4,4279
-biblicus/sources.py,sha256=EFy8-rQNLsyzz-98mH-z8gEHMYbqigcNFKLaR92KfDE,7241
+biblicus/retrieval.py,sha256=qAauHbnQcxtWZzonyOuwgSsffPyZ--0Z8wW-dEYk0z4,4287
+biblicus/sources.py,sha256=FNwW1FWts0jxWIL3AHon7D6c5ZatyG9AGFqzn1Id5mE,8504
 biblicus/time.py,sha256=3BSKOSo7R10K-0Dzrbdtl3fh5_yShTYqfdlKvvdkx7M,485
 biblicus/uris.py,sha256=xXD77lqsT9NxbyzI1spX9Y5a3-U6sLYMnpeSAV7g-nM,2013
 biblicus/user_config.py,sha256=UXUYBNUN4FR37ggZGJG1wv3K8XzsMR8pXW1T18lrivw,6495
@@ -43,13 +43,18 @@ biblicus/analysis/schema.py,sha256=MCiAQJmijVk8iM8rOUYbzyaDwsMR-Oo86iZU5NCbDMM,4
 biblicus/analysis/topic_modeling.py,sha256=mNBiRMpY5Jtyz8Aj-WXYY8guEghx9jozTfgveinJLoc,22135
 biblicus/backends/__init__.py,sha256=WJSvXc6boEj8PeFr__AC6l_0lfBPJpaVgMbVq30vtZU,1669
 biblicus/backends/base.py,sha256=Erfj9dXg0nkRKnEcNjHR9_0Ddb2B1NvbmRksVm_g1dU,1776
-biblicus/backends/embedding_index_common.py,sha256=WdmnOr8-QaLZH9b7TNmiWg_5auIqDYviXc-twOd7WsM,10216
-biblicus/backends/embedding_index_file.py,sha256=igwlWHuqQD72qyJ3OZ9uyBGeha7AMg6uA3Uu9QD_2_M,9552
-biblicus/backends/embedding_index_inmemory.py,sha256=oZCNS1kMajEj8u9C5iOjs_dAwhZ8ajU_br0F_8onRLY,9457
-biblicus/backends/hybrid.py,sha256=_kqpvD9V6oFdFSXE7K_rTN8qqH-vDIigPDX0uQxFhaM,10828
-biblicus/backends/scan.py,sha256=hdNnQWqi5IH6j95w30BZHxLJ0W9PTaOkqfWJuxCCEMI,12478
+biblicus/backends/embedding_index_common.py,sha256=wwvp6DjcaAwq-cp2jaO9TvnxDM7JDi-kpgT9uQG9Cxs,11552
+biblicus/backends/embedding_index_file.py,sha256=vibYEWa12Gx-Pm8WnuBnMfBaKiwlAvVW1dEzWJc6JO4,9856
+biblicus/backends/embedding_index_inmemory.py,sha256=LYiNBRmnh4DB8hmlBxMrm_uNmWi46Jt2EvjCuJGm2DI,9711
+biblicus/backends/hybrid.py,sha256=vlsN9N6FZ5A3dQtGXy0W89L4qNQX5EYJNvUuj2-Uqaw,10897
+biblicus/backends/scan.py,sha256=NBlfFHkDS3vdv70bgggK-jHykQC3W_i-RDaa97LEwKE,12548
 biblicus/backends/sqlite_full_text_search.py,sha256=tkFYdKwH6WvAF3En1fvGN_03Ud0_Z1igGxhUW4meCbA,24496
-biblicus/backends/tf_vector.py,sha256=HFV2aj1i4uxFYBd4GOv2M665ZcaQNgHWA4qmLbpIc4k,15220
+biblicus/backends/tf_vector.py,sha256=Z5MiEpbZ7A4UtRLYPEU1g8ubjWV5vuyPG40FpElEVzA,15119
+biblicus/context_engine/__init__.py,sha256=cIJWTUwOewW1x13a2n0YKfr4-XU0IwlVdAH_0pckfKk,1337
+biblicus/context_engine/assembler.py,sha256=ot5mdGJTA1nO8uUP_J_yGXgfVqQhFuEQJ3BH-HF4ZaY,42336
+biblicus/context_engine/compaction.py,sha256=2bLaCpT48d1TL7vt9rrcRCgfdHeWWp9LX85Cgij12o0,2921
+biblicus/context_engine/models.py,sha256=jesVd83ZQcatO-7yNlzwKkactSQ-e1znYuWof4rxVFg,12762
+biblicus/context_engine/retrieval.py,sha256=au_mN8VYc_MhIlbMGHfDf2IK0UWAigj7R5NFXFZ0Kz8,4143
 biblicus/extractors/__init__.py,sha256=ci3oldbdQZ8meAfHccM48CqQtZsPSRg3HkPrBSZF15M,2673
 biblicus/extractors/base.py,sha256=ka-nz_1zHPr4TS9sU4JfOoY-PJh7lbHPBOEBrbQFGSc,2171
 biblicus/extractors/deepgram_stt.py,sha256=VI71i4lbE-EFHcvpNcCPRpT8z7A5IuaSrT1UaPyZ8UY,6323
@@ -71,16 +76,16 @@ biblicus/extractors/unstructured_text.py,sha256=l2S_wD_htu7ZHoJQNQtP-kGlEgOeKV_w
 biblicus/text/__init__.py,sha256=MiaGAY7xWlUCeBzDzNz6pJnSMiU_Ge5EmlSiEzhqTRo,947
 biblicus/text/annotate.py,sha256=asmpj3_s_t8hl6stEg99apmqxAhDTkoPzHhZNggYE3Y,8355
 biblicus/text/extract.py,sha256=pdnUiZWtfCUj7kZK5zhd-tjqokgmhYYheWhyN3iShRU,7669
-biblicus/text/link.py,sha256=Xl0yxD1rvbRJRVdWG_ZP6zgmbpgWSJYcUcNM06-OUWU,20077
+biblicus/text/link.py,sha256=2IdOi3WgyBKPFau0bpS1eToV1q2v_6wq5RK5_P_qUDg,20448
 biblicus/text/markup.py,sha256=8jj9aX03HiZTOWdPs_VC4JLpQ7TlPHgGuXj_QUQIHVw,6265
 biblicus/text/models.py,sha256=REp6RowUWFdV-6y437JENP7XtGKt57BOvVtF91KmUqI,10853
-biblicus/text/prompts.py,sha256=Z5fSsy1Xzr0rCI0WZ3djiEQlbRDncyNBQ7_ZoWVPL4g,6704
+biblicus/text/prompts.py,sha256=-M-8sQ7Dfm1k4j6Kn4ekAuiYe_TkIwLu2VSgpas9rUU,6881
 biblicus/text/redact.py,sha256=tkDRmA0VvOZwMryEmBPLEHf3Z6VHJkkaWjBaNIMyGZ0,8415
 biblicus/text/slice.py,sha256=dlHxGO8c5P8BszXGwlNQoQ-cyWjJf6PfS1LUBJXXGEE,5762
 biblicus/text/tool_loop.py,sha256=w1PGLBvIemOdi6l0ArdYDVL7zgx-RC76bBOO0PKqpt0,11831
-biblicus-0.16.0.dist-info/licenses/LICENSE,sha256=lw44GXFG_Q0fS8m5VoEvv_xtdBXK26pBcbSPUCXee_Q,1078
-biblicus-0.16.0.dist-info/METADATA,sha256=D_qXRDdM_7LnOD-GBtnqwldEWOKjt_uV-TfqvnL8MAU,30933
-biblicus-0.16.0.dist-info/WHEEL,sha256=wUyA8OaulRlbfwMtmQsvNngGrxQHAvkKcvRmdizlJi0,92
-biblicus-0.16.0.dist-info/entry_points.txt,sha256=BZmO4H8Uz00fyi1RAFryOCGfZgX7eHWkY2NE-G54U5A,47
-biblicus-0.16.0.dist-info/top_level.txt,sha256=sUD_XVZwDxZ29-FBv1MknTGh4mgDXznGuP28KJY_WKc,9
-biblicus-0.16.0.dist-info/RECORD,,
+biblicus-1.0.0.dist-info/licenses/LICENSE,sha256=lw44GXFG_Q0fS8m5VoEvv_xtdBXK26pBcbSPUCXee_Q,1078
+biblicus-1.0.0.dist-info/METADATA,sha256=oyWd6igX6I3o46-VjOAUVskj1pLzZ8DovsTV1mqpPoY,30940
+biblicus-1.0.0.dist-info/WHEEL,sha256=wUyA8OaulRlbfwMtmQsvNngGrxQHAvkKcvRmdizlJi0,92
+biblicus-1.0.0.dist-info/entry_points.txt,sha256=BZmO4H8Uz00fyi1RAFryOCGfZgX7eHWkY2NE-G54U5A,47
+biblicus-1.0.0.dist-info/top_level.txt,sha256=sUD_XVZwDxZ29-FBv1MknTGh4mgDXznGuP28KJY_WKc,9
+biblicus-1.0.0.dist-info/RECORD,,