Text to Data: Make Your Code Malleable, Not Brittle
summary
A concrete extractor that tags text then runs a three‑step pipeline of tuple relational calculus queries to select, transform and project locations is available as the Python package kante and runs from a Jupyter notebook. Examples show email and date extraction from messy headers, table extraction from PDF text into a list of dicts via get_table, and functions like in_between plus user defined predicates to remove duplicates The demos work end to end and reduce regex sprawl into declarative queries.
AI-written summary · about
chapters (9)
- 0:00Text to data extraction
- 1:09Tagging text with annotations
- 3:38Querying tagged locations
- 8:57Adjusting location offsets
- 10:02Demo setup
- 11:05Extracting emails from headers
- 17:06Extracting tables from bills
- 20:40Declarative extraction advantages
- 22:51Q&A
related talks
44mWorkshop 3: Data Engineering - A Roadmap... by Ramon Perez [Part 2]
18mAutomate Your City Data with Python
35mBad data - anecdotes and examples from the real world
24mExtracting data from Excel with Python
49mThe CV: A Data Scientist's View
31mHoly D@t*! How to Deal with Imperfect, Unclean Datasets
50mBuilding Data Science Workflows with Kotlin
28mYour data, your insights: creating personal data projects to (re-)own the...