Submits identifiers (gene symbols, UniProt/ENSEMBL accessions, etc.) to the
Reactome AnalysisService and returns the most enriched pathways with their
statistics. Reuse the returned `token` with `get_analysis_results` to
page/sort/filter without re-submitting.

Args:
    identifiers: A list, or whitespace/comma-separated string, of
        identifiers (e.g. "TP53 EGFR BRCA1 MDM2 CDKN1A").
    projection: If True (default), project non-human identifiers onto human
        pathways (uses `/identifiers/projection`).
    species: Optional species filter for the results.
    sort_by: One of ENTITIES_PVALUE, ENTITIES_FDR, ENTITIES_RATIO, etc.
    p_value: Keep pathways with entity p-value <= this (default 1.0 = all).
    page_size: Pathways per page (default 20).
    page: 1-based page number.
    include_interactors: Expand the analysis with IntAct interactors.

Returns:
    Dict with `token`, `pathwaysFound`, `identifiersNotFound`, and
    `pathways` (each `{stId, name, species, entitiesFound, entitiesTotal,
    pValue, fdr, reactionsFound, reactionsTotal}`), plus resource/species
    summaries and any `warnings`.

Use this for any non-trivial payload (full SBML/SBGN, PDF documents, large
diagram images) instead of pulling it into the chat.

Args:
    stable_id: Event/diagram/reaction stable id; for `kind="fireworks"` a
        species name or taxId.
    kind: One of "event" (sbml/sbgn), "diagram" (png/jpg/gif/svg),
        "document" (pdf), "reaction" (png/jpg/gif/svg), or "fireworks"
        (png/jpg/gif/svg, species-level overview).
    ext: File extension valid for the chosen `kind`.
    save_dir: Local directory to save into; created if absent.

Returns:
    Absolute path to the saved file, or a string starting with "ERROR:".

The inverse of `list_pathway_events`: given a molecule/complex, which
pathways involve it?

Args:
    stable_id: PhysicalEntity stable id, e.g. "R-HSA-199420".
    species: Optional species name/taxId filter.
    all_forms: If True, also match every other form of the entity (e.g.
        phosphorylated/cleaved variants sharing a ReferenceEntity).

Returns:
    Dict `{count, pathways}` of Pathway records that contain the entity.

Args:
    token: The `token` returned by `analyze_identifiers`.
    page: Zero-based page number.
    page_size: Identifiers per page.

Returns:
    Dict `{count, notFound}` of the unmapped identifiers.

Args:
    token: The `token` returned by `analyze_identifiers`.
    species: Optional species filter.
    sort_by: Sort key (e.g. ENTITIES_PVALUE, ENTITIES_FDR).
    p_value: Keep pathways with entity p-value <= this.
    page: 1-based page number.
    page_size: Pathways per page.
    resource: Identifier resource to score against (default "TOTAL").

Returns:
    Same shape as `analyze_identifiers`.

Args:
    stable_id: Complex stable id, e.g. "R-HSA-83538".
    exclude_structures: If True, omit structural sub-complexes.

Returns:
    Dict `{count, subunits}` of the constituent PhysicalEntities.

Args:
    stable_ids: A list of identifiers, or a comma/whitespace-separated
        string (e.g. "R-HSA-69278, R-HSA-69620").

Returns:
    Dict `{count, entries}` where `entries` is the list of matched records.

Works for pathways, reactions, physical entities, complexes, regulators,
etc. Use `attribute` to retrieve a single field cheaply, or `enhanced` to
include indirect references.

Args:
    stable_id: Reactome stable id (e.g. "R-HSA-69278") or numeric dbId.
    enhanced: If True, query the `/data/query/enhanced/{id}` variant.
    attribute: If set, fetch only this attribute (e.g. "displayName",
        "speciesName", "stIdVersion").
    max_chars: Size guard. If the JSON exceeds this, large nested fields
        are replaced with a summary (default 50000).

Returns:
    The object's JSON record (name, displayName, schemaClass, definition,
    species, compartments, literatureReferences, etc.), size-guarded; or
    `{stId, attribute, value}` when `attribute` is given.

Useful for building breadcrumbs / locating where a reaction sits.

Args:
    stable_id: Event (pathway or reaction) stable id, e.g. "R-HSA-69620".

Returns:
    Dict `{count, ancestors}` where each ancestor is an ordered path of
    events from the queried event to a top-level pathway.

Args:
    stable_id: Event stable id, e.g. "R-HSA-69620".

Returns:
    Dict `{count, participants}` grouping each participating PhysicalEntity
    with its ReferenceEntities (UniProt/ChEBI/etc. cross-references).

SBML connects Reactome straight into systems-biology tooling (and the
BioModels workflow). Large exports are head/tail truncated with a pointer
to `download_export` so multi-MB XML never floods the chat.

Args:
    stable_id: Event stable id, e.g. "R-HSA-69620".
    fmt: "sbml" (default) or "sbgn".
    max_chars: Max characters to return inline (default 50000).

Returns:
    Dict with `stId`, `format`, `mimeType`, `total_chars`, `truncated`, and
    either `content` (full text) or `content_head` + `content_tail` +
    `note` when truncated.

Args:
    accession: A UniProt/ChEBI accession, e.g. "P04637".
    page: 1-based page number; -1 (default) returns all.
    page_size: Page size; -1 (default) returns all.

Returns:
    The interactor summary object, including the list of interacting
    accessions and their interaction scores.

Returns:
    Dict `{count, diseases}` of Disease records (Disease Ontology terms).

Recurses the full sub-tree, so the result is the flattened set of events
beneath the given pathway.

Args:
    stable_id: Pathway stable id, e.g. "R-HSA-69278".

Returns:
    Dict `{count, events}` of Event records (Pathways + ReactionLikeEvents).

Args:
    main_only: If True (default), return only the main curated species;
        otherwise return every species including computational inferences.

Returns:
    Dict `{count, species}` of Species records (name, taxId, abbreviation).

Args:
    species: Species name ("Homo sapiens") or NCBI taxId ("9606").

Returns:
    Dict `{count, pathways}` of top-level Pathway records.

Matches pathways, reactions, physical entities, and reference molecules by
name, synonym, identifier, and cross-reference.

Args:
    query: Search text (e.g. "TP53", "glycolysis", "apoptosis").
    species: Optional species filter, e.g. "Homo sapiens".
    types: Optional comma-separated type filter, e.g. "Pathway,Reaction"
        or "Protein".
    rows: Max entries to return per type group (1-100).
    start: Zero-based offset into the result set for pagination.
    cluster: Group hits by type (recommended; keeps the response compact).

Returns:
    Dict with `entries` (flattened `{stId, name, type, exactType, species,
    referenceIdentifier, compartmentNames, databaseName, ...}` with HTML
    highlight markup stripped), `entriesCount` (total matches), and
    `resultTypes` (per-type match counts).