scitex_hub

SciTeX Hub - CLI tools and APIs for SciTeX deployment and management.

Usage:

pip install scitex-hub scitex-hub –help

Python API:

>>> import scitex_hub
>>> client = scitex_hub.CloudClient()
>>> client.scholar_search("neural networks")
>>> client.enrich_bibtex("@article{...}")

MCP Server:

scitex-hub serve # stdio (Claude Desktop) scitex-hub serve -t sse # SSE (remote)

Functions

`eval_js`(code[, timeout])	Evaluate JavaScript in the user's connected browser.
`get_context`([page])	Get web app context: username, page, skills, available actions.
`get_version`()	Get scitex-hub version.
`health_check`([endpoint])	Check scitex-hub service health.
`ui_action`(steps[, delay_ms])	Drive browser UI: navigate, highlight, click, fill, scroll.

scitex_hub.get_version() → str[source]: Get scitex-hub version.

scitex_hub.health_check(endpoint: str | None = None) → dict[source]

Check scitex-hub service health.

Parameters:: endpoint (str, optional) – HTTP endpoint to check. If None, returns local package info.
Returns:: Health status with version, environment, and service status.
Return type:: dict

scitex_hub.get_context(page: str = '', **kw) → dict[source]

Get web app context: username, page, skills, available actions.

Convenience wrapper over CloudClient.get_context.

Parameters:

page (str, optional) – Page name or URL fragment to query context for. "" means the current page.
**kw – Forwarded to CloudClient (e.g. base URL, auth credentials).

Returns:

Context dict (username, page, actions, …).

Return type:

dict

scitex_hub.eval_js(code: str, timeout: int = 10, **kw) → dict[source]

Evaluate JavaScript in the user’s connected browser.

Convenience wrapper over CloudClient.eval_js.

Parameters:

code (str) – JavaScript code to evaluate.
timeout (int, optional) – Seconds to wait for the result (default 10).
**kw – Forwarded to CloudClient.

Returns:

Result of the JavaScript evaluation.

Return type:

dict

scitex_hub.ui_action(steps: list, delay_ms: int = 900, **kw) → dict[source]

Drive browser UI: navigate, highlight, click, fill, scroll.

Convenience wrapper over CloudClient.ui_action.

Parameters:

steps (list) – List of action-step dicts; each describes one browser action.
delay_ms (int, optional) – Milliseconds to wait between steps (default 900).
**kw – Forwarded to CloudClient.

Returns:

Result summary.

Return type:

dict

class scitex_hub.CloudClient(api_key: str | None = None, base_url: str | None = None)[source]

Python client for SciTeX Hub API.

Parameters:

api_key (str, optional) – API key for authenticated endpoints. Falls back to SCITEX_HUB_API_KEY env var.
base_url (str, optional) – Cloud server URL. Falls back to SCITEX_HUB_URL env var or https://scitex.ai.

Examples

>>> client = CloudClient()
>>> result = client.scholar_search("machine learning")
>>> print(result["papers"])

__init__(api_key: str | None = None, base_url: str | None = None)[source]

_request(method: str, endpoint: str, data: dict | None = None, files: dict | None = None, auth_required: bool = True) → dict[source]: Make HTTP request to SciTeX Hub API.

scholar_search(query: str, limit: int = 10) → dict[source]

Search papers via SciTeX Hub (public, no auth required).

Parameters:

query (str) – Search query string.
limit (int, optional) – Maximum number of results.

Returns:

Search results with papers list.

Return type:

dict

crossref_search(query: str, rows: int = 10, offset: int = 0) → dict[source]

Search CrossRef database via cloud proxy.

Parameters:

query (str) – Search query string.
rows (int, optional) – Number of results per page.
offset (int, optional) – Offset for pagination.

Returns:

CrossRef search results.

Return type:

dict

crossref_by_doi(doi: str) → dict[source]

Get CrossRef metadata by DOI.

Parameters:: doi (str) – Digital Object Identifier.
Returns:: CrossRef metadata for the DOI.
Return type:: dict

enrich_bibtex(bibtex_content: str, use_cache: bool = True, timeout: int = 120) → str[source]

Enrich BibTeX content with metadata.

Parameters:

bibtex_content (str) – BibTeX content to enrich.
use_cache (bool, optional) – Whether to use cached results.
timeout (int, optional) – Maximum time to wait for enrichment (seconds).

Returns:

Enriched BibTeX content.

Return type:

str

writer_compile(project_id: str, document_type: str = 'manuscript') → dict[source]

Compile LaTeX manuscript via cloud.

Parameters:

project_id (str) – Project identifier.
document_type (str, optional) – Type of document: manuscript, supplementary, revision.

Returns:

Compilation result with PDF URL.

Return type:

dict

project_list_files(project_id: str, path: str = '') → dict[source]

List files in a cloud project.

Parameters:

project_id (str) – Project identifier.
path (str, optional) – Subdirectory path.

Returns:

File listing.

Return type:

dict

get_context(page: str = '') → dict[source]

Get web app context for AI agents.

Parameters:: page (str, optional) – Current page URL (e.g. /writer/).
Returns:: Context including username, active skill, all skills, available actions, and aggregated context.
Return type:: dict

eval_js(code: str, timeout: int = 10) → dict[source]

Evaluate JavaScript in the user’s browser.

Parameters:

code (str) – JavaScript code to evaluate.
timeout (int, optional) – Max seconds to wait for result.

Returns:

Evaluation result from the browser.

Return type:

dict

ui_action(steps: list, delay_ms: int = 900) → dict[source]

Drive browser UI actions via WebSocket relay.

Parameters:

steps (list) – Action steps (navigate, highlight, click, fill, scroll, clear).
delay_ms (int, optional) – Delay between steps in milliseconds.

Returns:

Result with number of steps sent.

Return type:

dict

status() → dict[source]

Check cloud service status.

Returns:: Status information including cloud availability.
Return type:: dict

class scitex_hub.Environment(name: str, docker_compose_file: str, env_file: str, host: str, port: int, description: str)[source]

Deployment environment configuration.

name: str

docker_compose_file: str

env_file: str

host: str

port: int

description: str

property env_path: Path: Get full path to .env file.

property compose_path: Path: Get full path to docker-compose file.

__init__(name: str, docker_compose_file: str, env_file: str, host: str, port: int, description: str) → None

scitex_hub.get_environment(name: str | None = None) → Environment[source]: Get environment configuration by name or auto-detect.

class scitex_hub.DockerManager(env: Environment | None = None, project_root: Path | None = None)[source]

Manage Docker containers for SciTeX Hub.

__init__(env: Environment | None = None, project_root: Path | None = None)[source]

_find_project_root() → Path[source]: Find project root by looking for pyproject.toml.

_run_compose(args: List[str], capture: bool = False) → CompletedProcess[source]: Run docker-compose command.

build(no_cache: bool = False) → int[source]: Build Docker containers.

up(detach: bool = True) → int[source]: Start Docker containers.

down(volumes: bool = False) → int[source]: Stop Docker containers.

restart() → int[source]: Restart Docker containers.

logs(follow: bool = False, tail: int | None = None, service: str | None = None) → int[source]: Show container logs.

ps() → int[source]: Show container status.

status() → dict[source]: Get detailed container status.

Modules

`appmaker`	SciTeX App Tools — init, validate, develop, publish, and manage app plugins.
`module`	SciTeX Hub Module -- decorator and output APIs for custom workspace modules.
`project`	SciTeX Hub project management.
`sdk`	SciTeX SDK — backward-compatible re-export from scitex-app.