On-demand workflow combination and variant generation

A system for workflow recombination, comprising:

• a workflow database;
• a domain knowledgebase;
• and
• a processor programmed to:
• receive a query from a client device indicating a topic for crawled workflows;
• crawl one or more network resources to identify the crawled workflows, including to initiate the crawl using a seed list of web addresses, the seed list including web document resources that relate to the topic to be crawled, to identify additional web addresses in the web document resources, and to add the additional web addresses to the one or more network resources to visit to retrieve the crawled workflows;
• crawl the one or more network resources to identify domain terminology so that terms and their relations are automatically extracted, refined, and organized;
• add the domain terminology to a domain knowledgebase;
• add the crawled workflows to the workflow database including previously stored workflows;
• annotate the crawled workflows stored in the workflow database utilizing the domain terminology as standardized terms to include in the annotations, each of the crawled workflows including an ordered sequence of steps, each of the steps indicating an action to be performed described using the terms stored in the domain knowledgebase, the annotations indicating pre-conditions to be fulfilled for execution of the respective steps to be performed and outputs of the respective steps, the pre-conditions and outputs being specified in the annotations as standardized terms from the domain knowledgebase;
• extract sub-workflows from the crawled workflows by splitting the crawled workflows into sub-workflows at annotations that exist between the steps, a first of the sub-workflows including the respective annotation as an output of the first of the sub-workflows, a second of the sub-workflows including the respective annotation as an input to the second of the sub-workflows;
• identify annotations in common between the crawled sub-workflows as extracted and sub-workflows of the previously stored workflows, including identifying beginning sub-workflows that include the respective annotation as an output and identifying ending sub-workflows that include the respective annotation as an input;
• and
• recombine the crawled sub-workflows and the sub-workflows of the previously stored workflows into new workflows by splicing together the sub-workflows at the annotations in common, such that each new workflow is formed from one of the beginning sub-workflows followed by one of the ending sub-workflows, the recombining creating new workflows using portions of the sub-workflows of the crawled workflows and portions of the sub-workflows of the previously stored workflows.

A non-transitory computer-readable medium comprising instructions for workflow recombination that, when executed by a processor, cause the processor to:

• receive a query from a client device indicating a topic for crawled workflows;
• crawl one or more network resources to identify the crawled workflows, including to initiate the crawl using a seed list of web addresses, the seed list including web document resources that relate to the topic to be crawled, to identify additional web addresses in the web document resources, and to add the additional web addresses to the one or more network resources to visit to retrieve the crawled workflows;
• crawl the one or more network resources to identify domain terminology so that terms and their relations are automatically extracted, refined, and organized;
• add the domain terminology to a domain knowledgebase;
• add the crawled workflows to a workflow database including previously stored workflows;
• annotate the crawled workflows stored in the workflow database utilizing the domain terminology as standardized terms to include in the annotations, each of the workflows including an ordered sequence of steps, each of the steps indicating an action to be performed described using the terms stored in the domain knowledgebase, the annotations indicating pre-conditions to be fulfilled for execution of the respective steps to be performed and outputs of the respective steps, the pre-conditions and outputs being specified in the annotations as standardized terms from the domain knowledgebase;
• extract sub-workflows from the crawled workflows by splitting the crawled workflows into sub-workflows at annotations that exist between the steps, a first of the sub-workflows including the respective annotation as an output of the first of the sub-workflows, a second of the sub-workflows including the respective annotation as an input to the second of the sub-workflows;
• identify annotations in common between the crawled sub-workflows as extracted and sub-workflows of the previously stored workflows, including identifying beginning sub-workflows that include the respective annotation as an output and identifying ending sub-workflows that include the respective annotation as an input;
• and
• recombine the crawled sub-workflows and the sub-workflows of the previously stored workflows into new workflows by splicing together the sub-workflows at the annotations in common, such that each new workflow is formed from one of the beginning sub-workflows followed by one of the ending sub-workflows, the recombining creating new workflows using portions of the sub-workflows of the crawled workflows and portions of the sub-workflows of the previously stored workflows.