synamine.discover¶

Process discovery algorithms.

discover ¶

High-level discovery API -- delegates to algorithm modules.

discover_dfg ¶

discover_dfg(
    log: EventLog, *, noise_threshold: float = 0.0
) -> DirectlyFollowsGraph

Discover a Directly-Follows Graph from an event log.

PARAMETER	DESCRIPTION
`log`	The event log to analyze. TYPE: `EventLog`
`noise_threshold`	Edges with frequency below this proportion of the max are removed. TYPE: `float` DEFAULT: `0.0`

RETURNS	DESCRIPTION
`DirectlyFollowsGraph`	The discovered DFG with edge frequencies.

discover_variant_trie ¶

discover_variant_trie(
    log: EventLog, *, min_count: int = 1
) -> VariantTrie

Build a variant trie (prefix tree) from an event log.

The trie represents all trace variants as a tree structure where shared prefixes are merged. Designed for dendrogram visualization.

PARAMETER	DESCRIPTION
`log`	The event log to analyze. TYPE: `EventLog`
`min_count`	Prune variants with fewer than this many traces. TYPE: `int` DEFAULT: `1`

RETURNS	DESCRIPTION
`VariantTrie`	The variant prefix tree.

discover_petri_net ¶

discover_petri_net(
    log: EventLog,
    *,
    algorithm: str = "alpha",
    noise_threshold: float = 0.0,
    dependency_threshold: float = 0.5,
    and_threshold: float = 0.65,
    loop_two_threshold: float = 0.5,
) -> PetriNet

Discover a Petri Net from an event log.

PARAMETER	DESCRIPTION
`log`	The event log to analyze. TYPE: `EventLog`
`algorithm`	Discovery algorithm to use ("alpha" or "heuristic"). TYPE: `str` DEFAULT: `"alpha"`
`noise_threshold`	Filter infrequent edges (0.0 = keep all). TYPE: `float` DEFAULT: `0.0`
`dependency_threshold`	Heuristic Miner: minimum dependency measure to keep an edge. TYPE: `float` DEFAULT: `0.5`
`and_threshold`	Heuristic Miner: threshold for AND-split/join detection. TYPE: `float` DEFAULT: `0.65`
`loop_two_threshold`	Heuristic Miner: threshold for length-2 loop detection. TYPE: `float` DEFAULT: `0.5`

RETURNS	DESCRIPTION
`PetriNet`

discover_process_tree ¶

discover_process_tree(
    log: EventLog, *, noise_threshold: float = 0.0
) -> ProcessTree

Discover a Process Tree from an event log using the Inductive Miner.

PARAMETER	DESCRIPTION
`log`	The event log to analyze. TYPE: `EventLog`
`noise_threshold`	Filter infrequent edges (0.0 = keep all, 0.2 = remove bottom 20%). TYPE: `float` DEFAULT: `0.0`

RETURNS	DESCRIPTION
`ProcessTree`

convert_dfg_to_petri_net ¶

convert_dfg_to_petri_net(
    dfg: DirectlyFollowsGraph,
) -> PetriNet

Convert a Directly-Follows Graph to a Petri Net.

PARAMETER	DESCRIPTION
`dfg`	The DFG to convert. TYPE: `DirectlyFollowsGraph`

RETURNS	DESCRIPTION
`PetriNet`	A Petri Net with one transition per activity.