Master 2.5 GB of unstructured specification documents with ease

In this talk, we present lessons learned from and practical advice on how to deal with a large body of specification documents in your next project. We introduce our approach as well as code excerpts from our powerful toolset to transform a large set of unstructured and partially corrupt specification documents into structured JSON Files. Finally, we showcase a simple, yet powerful Javascript frontend which requires no additional infrastructure to present the compiled artefacts in an intuitive and responsive user interface. In particular this talk covers the following topics: How to make use of pywin32 to access layout and content information from partially corrupt .doc and .docx files and create simple JSON files with UTF-8 encoding. Identify and categorize signal words in your specification. Use pandas to compile content based recommender functionality. Use networkx and py2cytoscape to visualize call sequences and semantic relationships in your specification. Present the compiled artefacts and identified relationships in an easy-to-use and lightweight Javascript browser interface without any additional infrastructure (i.e. no webserver and no database server).