I'm starting a project that will be done entirely through Python, except perhaps at the end for visualisation. So far I've been doing it with ArcGIS and arcpy, but we'd like to move it to an open source platform.
I've been searching and haven't been able to find out how exactly QGIS fits in with GDAL/OGR (and other open source projects like GRASS). It seems that QGIS includes GDAL/OGR but has lots of its own additions. I'm trying to figure out what's the easiest route to perform a lot of simple, high-level operations.
At the moment what I'm doing with arcpy is converting between rasters and points, extracting raster values into existing points, extracting tables and geometry to NumPy, hydrology, finding intersections and nearest distance between features. There is only one operation that I imagine would have to be reworked quite extensively (inserting points every x metres along a set of vectors).
Is the Python interface of QGIS sufficient to handle this, or is there an advantage in using GDAL/OGR and other packages? I'd like it to be as simple and readable as possible, as lots of other people will be using it and modifying it when we're finished.