Layout-agnostic complex document processing system

A system comprising:

• a storage service implemented by a first one or more electronic devices within a provider network;
• and
• a document processing service implemented by a second one or more electronic devices within the provider network, the document processing service including instructions that upon execution cause the document processing service to:
• receive a first request to analyze a document encoded by a file, the first request identifying a location provided by the storage service that stores the file;
• generate a job identifier (ID) of a job;
• transmit the job ID within a first response, wherein the first response is associated with the first request;
• obtain the file from the storage service;
• identify, within at least a portion of the document, one or more segments;
• determine, via use of one or more machine learning (ML) models, a classification of each of the one or more segments into one or more corresponding segment types;
• process each of the one or more segments using an analysis engine selected based the determined segment type for the segment to yield one or more results;
• receive a second request for an output of the job, the second request comprising the job ID;
• and
• transmit a second response including an output based on the one or more results wherein the second response is associated with the second request.

A computer-implemented method comprising:

• receiving a request to analyze a document encoded by a file;
• identifying, within at least a portion of the document, one or more segments;
• determining, via use of one or more machine learning (ML) models, a classification of each of the one or more segments into one or more corresponding segment types, wherein at least one segment type of the one or more corresponding segment types is a form;
• processing each of the one or more segments using an analysis engine selected based on the determined segment type for the segment to yield one or more results, the processing for each segment comprising:
• identifying a plurality of keys of the form and a plurality of values of the form;
• and
• determining a plurality of key-value pairings, wherein at least one of the plurality of key-value pairings comprises one of the plurality of keys and one of the plurality of values that corresponds to the one key;
• and
• transmitting a response including an output based on the one or more results.

A system comprising:

• a storage service implemented by a first one or more electronic devices within a provider network;
• and
• a document processing service implemented by a second one or more electronic devices within the provider network, the document processing service including instructions that upon execution cause the document processing service to:
• receive a request to analyze a document encoded by a file, the request identifying a location provided by the storage service that stores the file;
• obtain the file from the storage service;
• identify, within at least a portion of the document, one or more segments;
• determine, via use of one or more machine learning (ML) models, a classification of each of the one or more segments into one or more corresponding segment types, wherein at least one of the one or more corresponding segment types for a corresponding one of the segments is a form;
• process each of the one or more segments using an analysis engine selected based the determined segment type for the segment to yield one or more results, wherein to process each segment the document processing service is to:
• identify a plurality of keys of the form and a plurality of values of the form;
• and
• determine a plurality of key-value pairings, wherein at least one of the plurality of key-value pairings comprises one of the plurality of keys and one of the plurality of values that corresponds to the one key;
• and
• transmit a response including an output based on the one or more results.