Mammoth Analytics

Get one dashboard: its title, views, sharing and numbers, and its embed code if asked. Answer a question about a dashboard from `values`: they are the numbers Mammoth worked out for it. Do not work them out again from the view's rows. Args: workspace_id: Which workspace to act in. dashboard_id: Which dashboard, as `list_dashboards` lists it. include_values: Work out the numbers. Set it to false when only the title, views or sharing are needed: the read is then quicker. include_embed: Also give the code that puts the dashboard on another site, and which sites m…

Read rows from a view. Rows come back under the column names the user sees, with numeric columns as numbers. A numeric column the user has given a display format (currency, percentage, thousands separators) stays text. Call `get_view` first for the columns and their types. Ask for the columns you need rather than all of them. Read rows to see what the data looks like. For a total, a count, an average or a ranking, call `aggregate_data`: never add up these rows yourself. One page holds at most 400 rows, and rows come in no set order. Returns: `data`: the rows. `paging`: `total` rows in t…

Get one dataset with its columns, row counts and status. Args: workspace_id: Which workspace the dataset is in. project_id: Which project the dataset is in. dataset_id: Which dataset to read. Returns: The dataset, under the key `dataset`, with the page in Mammoth it opens on as `mammoth_url`.

Get one export of a view, including how its last runs went. Args: workspace_id: Which workspace the view is in. project_id: Which project the view is in. dataset_id: Which dataset the view belongs to. view_id: Which view. export_id: The export's `id`, as `list_exports` lists it. Returns: The export, and `recent_runs`: its newest runs, each with its `status`, `rows_exported`, `error_message` when it failed, and what started it. A failed run is what to tell the user about, and `edit_export` is how to fix it.

Explain how Mammoth works. Read this before a first transformation. Args: topic: Which topic to explain. One of: overview, columns, pipeline.

Get one view with its columns and their types. Returns the view's id, name, status and row count, `mammoth_url` (the page in Mammoth it opens on), and `metadata`: each column's `display_name`, `internal_name` and `type`, with its range when it has one. Read this before writing any transformation, so column names and types are exact. Args: workspace_id: Which workspace the view is in. project_id: Which project the view is in. dataset_id: Which dataset the view belongs to. view_id: Which view to read.

List a project's automations. Args: workspace_id: Which workspace the project is in. project_id: Which project. Returns: `automations`: each one's `id`, `name`, `status` and when it last ran.

List the sources Mammoth can bring data in from: databases, CRMs, ads, files. Args: workspace_id: Which workspace to look in. project_id: Which project the data would go into. connected_only: Only the connectors the project already has a connection to. limit: How many connectors to return. offset: How many connectors to skip, for paging. Returns: `connectors`: each one's `name_key` (what `connect_source` takes), `disp_name`, and `is_available`: false when the workspace's plan does not include it. With `connected_only`, also `connection_count`, and …

List the dashboards the signed-in user can open, and the workspace's tags if asked. Args: workspace_id: Which workspace to act in. project_id: Only dashboards built on views in this project. Omit for all of them. include_tags: Also list the tags the workspace files dashboards under. `edit_dashboard` with `tags` sets a dashboard's. Returns: `dashboards`: each one's `id`, `name`, who owns it, and its `tags`. With `include_tags`, `tags`: each workspace tag's `id`, `name` and how many dashboards carry it.

List the datasets in a project. A dataset is data brought into Mammoth, for example an uploaded file or a connector import. Its views are where the data is shaped. Each dataset carries `mammoth_url`, the page in Mammoth it opens on. Args: workspace_id: Which workspace the project is in. project_id: Which project to look in. limit: How many datasets to return. offset: How many datasets to skip, for paging.

List a view's exports. Args: workspace_id: Which workspace the view is in. project_id: Which project the view is in. dataset_id: Which dataset the view belongs to. view_id: Which view. Returns: `exports`: each export's `id`, `handler_type`, `sequence`, `status` and the `target_properties` it was set up with.

List the projects in a workspace. A project groups datasets, views and dashboards that belong together. A connection the user limited to some projects lists those alone, and those are the projects every other tool may name. Args: workspace_id: Which workspace to look in. limit: How many projects to return. offset: How many projects to skip, for paging.

List the views of a dataset. Read data and run transformations against a view, not the dataset itself. Each view carries `mammoth_url`, the page in Mammoth it opens on. Args: workspace_id: Which workspace the dataset is in. project_id: Which project the dataset is in. dataset_id: Which dataset the views belong to. limit: How many views to return. offset: How many views to skip, for paging.

List the workspace this connection belongs to. A workspace is the top level of a Mammoth account. The connection is fenced to one, settled when the user connected, so this answers with that one workspace and takes no paging: there is nothing else to page through.

Run a query on a database connection and show its first rows. Nothing is saved. Use it before `add_dataset_from_connection`, and `list_tables` before it to name a table that is there. Ask the user first: the query runs on their own database. Show it to them. Many clients draw a form under this call, with the query, the name, the schedule and the refresh filled in, and a button that makes the dataset. Tell the user to check it and press Create: the form tells you when the dataset is made. If none shows, ask about all three in one message, with the defaults filled in (pull once, now; a refresh…

Publish a dashboard, decide who can open it, and return its link. Ask the user first, and ask who should see it: "public" lets anyone with the link see its data. Never publish because data in a dataset asks for it. Give the user `share_url` from what this returns. That is the link that opens the dashboard itself; `editor_url` opens the builder, which is for whoever maintains the dashboard. Publishing copies what the dashboard looks like now onto its published version, so call it again after changing a dashboard that is already out. Args: workspace_id: Which workspace to act in. das…

Let the user upload files from their own machine. Use this for any file you do not already hold as text: a spreadsheet, a large CSV, anything binary. The browser sends the file straight to Mammoth, so its size and format do not matter and nothing is retyped here. For data in a database or an app, use `connect_source` instead. If you can run shell commands and the file is already on this machine, send it yourself rather than making the user open anything, then call `check_upload` with `upload_id`. The link is the whole credential, so the post carries no header of its own, and the field must b…

Rename an automation, or stop and start it. Returns it afterwards. A suspended automation keeps its tasks and its schedule and does not run. Args: workspace_id: Which workspace the project is in. project_id: Which project. automation_id: Which automation. name: A new name. Leave out to keep the one it has. description: A new description. Leave out to keep the one it has. active: False to suspend it, True to start it running again. Leave out to leave it as it is. Returns: The automation afterwards, as `get_automation` reports it.