These are terms that Shesmu uses with very specific meanings.
A unit of work for Shesmu to do. A plugin provides a definition that determines what parameters an action needs, an olive fills in those parameters, and then Shesmu will perform an action until it succeeds. An action is defined by its parameters; that is, actions with the same parameters are duplicates. This allows Shesmu to deduplicate actions across runs with no state.
A Prometheus alert. Shesmu follows the semantics laid out by Prometheus: an alert is effectively a set of key-value string pairs that define some emergency for a limited (but extendible) period of time.
An algebraic data type is a type that can be in one of several states. It can do some of the job of inheritance in object-oriented languages. For details see the guide to algebraic data types.
Syntax in an olive that reshapes data. In a functional programming model, these
are filter (Where), map (Let), or reduce (Group).
A value that is available to an olive. Constants are not necessarily constant
(e.g., now is a constant for the current time). A few constants are
built-in, but most are provided by plugins or other olives.
An encoded representation of a Shesmu type. Shesmu types are different from Java types. The encoded form is guaranteed to be a valid Java identifier. Internally, these types are called imyhat, which is Ancient Egyptian for “mould or form”.
A predicate that determines which actions are relevant. This is used in the UI and REST interface to select subsets of actions. Filters may describe properties of the action itself or the olives that generated it.
A function that is available to an olive. Like constants, a few functions are built-in, but most are provided by plugins or other olives.
A set of variables that should be treated as a group. In many data formats,
there are common grouping conditions. Gangs are an easy to pre-define these
variable groups. They can be used in Group By conditions, converted into
strings, or converted into tuples.
The data that is available to olives. This data is effectively tabular even though it is never really encoded that way. Each record/row gets processed by an olive where each column is available as a variable. Columns cannot be missing/null. Missing entries need to either be substituted with dummy values or optional values. If the schema is ragged, it can be expressed as a JSON value.
A homogeneous collection of items. Shesmu does not respect order on collections and automatically removes duplicate items. This is an intentional design to choice to prevent matched lists where the indices mean something. Create a list of tuples in this case.
Shesmu’s objects are not like Java’s object. They are more like JSON objects or tuples with property names instead of indices.
A Shesmu program element that takes the original input data and manipulates it down to one of the following goals:
Each olive independently handles its data (that is, no other olive in the file/script can affect this olive). Files/scripts are scheduled as a unit, so if one olive exceeds the maximum runtime, other olives in the file may not run.
Shesmu does not allow nulls. Instead, it uses option types implemented using Java’s Optional.
Only types explicitly declared as optional with a ? (e.g., integer?) can
be absent. Optionals cannot nest (i.e., integer?? is invalid).
An export from an olive to a database. A refiller is meant to empty and replace the contents of a database with the current set of data computed by an olive. This is not likely the correct implementation of a refiller, but it is conceptually useful.
It can be useful to think of a refiller as performing an upsert operation on an entire table in a database.
A filter with a name that allows the UI to quickly select an interesting set of actions. They can be used for debugging or as dashboards.
A fingerprint (hash or value) derived from the values of variables read by an olive.
An action might need to be re-run if the input data changes. Since an olive doesn’t necessarily use all of the input data, it can be useful to generate a hash of only the data that was used by an olive in deciding to run an action. This signature can be used by the action to determine under what conditions it should be rerun. The input format determines which variables should make it into the signature and which should be excluded.
A method to stop actions from running. Since Shesmu may be talking to many systems at once, throttlers provide a way to take some of those systems down for maintenance or limit Shesmu’s traffic to them during an overload. Actions and olives check for permission to run from the throttlers and will quit early if the system is reported as overloaded.
An ordered, heterogeneous collection of values. Similar to Scala’s tuples.
A value from input data. This is the “column” that is being read.