Last 2 weeks:
- For UIST 2006 paper prototype
- Implemented field assignment (which no longer gets trapped in rich
text blobs and assigns too many fields)
- Generated scraper from template (which only scrapes one page).
- Supported more effective item selection using /.. trick.
- Added date parsing and parsing of numbers with comma separators.
- Fixed Longwell2 to recognize decimal numbers.
- Inserted "smart tags" into scraped items' field slots.
http://people.csail.mit.edu/dfhuynh/misc/autoscraper7.PNG
http://people.csail.mit.edu/dfhuynh/misc/autoscraper8.png
- Factored out auto scraper.
- Made auto scraper grab subsequent pages and also learn from some
subsequent pages.
- Made auto scraper grab innerHTML from scraped items, store them,
and subs them back.
- Changed highlighting mechanism to using translucent DIVs.
http://people.csail.mit.edu/dfhuynh/misc/autoscraper6.PNG
- Changed page links in Explore pane to resemble web site page link
style.
- Faded away web page except for scraped items.
http://people.csail.mit.edu/dfhuynh/misc/autoscraper9.png
- Linked up "smart tags" to Explore pane, clicking results in new
facets inserted into the pane, hovering highlights corresponding facets
- Started to implement range-based facets using trees
- Tried to fix the GRDDL bug in Piggy Bank
This week:
- UIST 2006 paper prototype
- Support range-based facets
- Support dual-highlighting for selection of items
- Support user's selection of paging links... not sure what to do
with it yet
- Persist auto scraper and reapply automatically when applicable
- Start design user study
- Start writing paper
David
Received on Mon Feb 27 2006 - 14:30:41 EST