Assign ``value`` to ``field`` iff the template has that field. Returns True if set.

Build a realistic Zotero item dict for stubbing. citation_key defaults to None — the field is omitted from the data dict entirely so add-when

Build a minimal Zotero item dict.

Build a mock requests.Response for CrossRef API.

Format a Zotero item's metadata as markdown. Args: item: A Zotero item dictionary. include_abstract: Whether to include the

Close database connection.

Create a mock requests.Response for arXiv API.

Patch _get_write_client to return (read_zot, write_zot).

Setup Zotero environment for CLI commands.

Install a write-capable FakeZotero; return it.

Patch os.path functions so the file path appears valid.

Patch for web-only mode: same client for read and write.

Normalize a tag-filter argument into a list[str] for pyzotero. Accepts every shape we've seen clients produce: - None / empty

Patch fitz in sys.modules so 'import fitz' inside server functions works.

Create a ZoteroSemanticSearch instance. Args: config_path: Path to configuration file db_path: Optional path to Zotero datab

items: list of pyzotero-shaped dicts with top-level 'key' and 'data'.

Build a Zotero item dict with optional relations.

Parse a BibTeX string and return a list of structured entries. Each entry: ``{"entry_type": str, "citekey": str, "fields": dict}``. Applies `

Update the semantic search database with Zotero items. Args: force_full_rebuild: Whether to rebuild the entire database

Search for similar documents. Args: query_texts: List of query texts n_results: Number of results to return

Attempt to extract fulltext and source from the item's best attachment. Preference order: 1. ``.zotero-ft-cache`` (Zotero's own alrea

Normalize an ISBN string and validate the checksum. Accepts ISBN-10, ISBN-13, and prefixed/URL forms (isbn:, https://isbndb.com/...). Strips

Parse a text block as a figure/table caption. Args: text: Full text of a PDF text block Returns: {"label": "Figure 3",

Import tool modules lazily to avoid heavy startup cost.

Build a realistic Zotero book item dict for stubbing.

Split a BibTeX authors string on ' and ' and structure each name.

Process a batch of items. _failed_docs: optional list (passed by reference from update_database) that collects (doc_text, metadata, d

Detect candidate figure/table regions on a PDF page. Provides coordinate grounding for area annotations: instead of guessing normalized

Get database connection, creating if needed.

Convert one parsed BibTeX entry dict into a Zotero item dict. ``template_fn(item_type)`` must return a pyzotero-style item template for the g

Get all items in a specific Zotero collection. Args: collection_key: The collection key/ID limit: Maximum number of items to

Apply `Key: value` line edits to an Extra field value. Extra is treated as newline-separated lines; lines of the form "Key: value" are matche

Attach captions to regions by spatial scoring. Score components: - vertical proximity (hard cutoff at max_distance) - horizontal x-r

Create searchable text from a Zotero item. Args: item: Zotero item dictionary Returns: Combined tex

Disable network lookups in _try_attach_oa_pdf for tests that create DOIs.

Convert one CSL JSON object into a Zotero item dict.

Delete an item from the semantic search database.

Create a dummy file so code that checks os.path.exists passes.

Download a PDF from a URL and attach it to a Zotero item. The URL is fetched through ``_guarded_pdf_get`` (SSRF guard + manual redirect re-va

Build a fake requests.get that returns responses keyed by URL substring.

Return True when ``_extract_text_from_file`` can return useful text. PDF and HTML are handled by their dedicated extractors elsewhere. For

Filter noise, de-duplicate, and merge fragmented candidate regions. Composite figures often appear as multiple raster/vector fragments (

Return the injected fake path.

Accept a CSL JSON string, single object, or list; return a list of dicts.

Create a minimal zotero.sqlite with the given item keys.

Build extra_fields dict from attachment_info for format_item_result.

Build an ISO-ish date string from bibtex fields.

Insert *citation_key* on the @-line of the first BibTeX entry if missing. Only touches entries whose @-line is empty (``@article{``) or has a bar

Build a minimal Zotero item dict for testing.

Create a LocalZoteroReader-like object for path resolution tests. Uses the real _resolve_attachment_path (not the fake override).

Normalize the caller's ``collections`` argument and resolve every spec (key, name, or '/'-path) to a live collection key. Raises ValueError w

Remove HTML/XML tags from a string. Args: raw_html: String containing HTML content. collapse_whitespace: If True, collapse runs o

Return a minimal Zotero item template.

Build a ZoteroSemanticSearch without touching Chroma or pyzotero.

Split *text* into overlapping passages on natural boundaries. Pure function (no I/O, no model load) so it is unit-testable in isolation. Retu

Truncate using tiktoken cl100k_base (correct for OpenAI models).

Collection/idempotency flags shared by every `add` subcommand.