Improvemenrt

[phatal]

Hi! I'm try to scrap very much points with multi thread (100 instances of bootstrap). I'm using JS calling python bootstrap.py from shell with locations and accounts reading from csv. But get an issue with DB. When I run some instances at one time (doesn't matter 2 or 100) I gettin error about dividin by zero at 186 line
delay = float( float(config.delay) / float(len(cells)) )

Have any idea how I can fix it? actually may be you can add function to load data from files? Now it's looks like some sort of BDSM for me :D cause i'm absolutely zero with python I forced to use JS wrap to manage all calls and data.

I'm just have about 4.9mln rect cells to scan with ~1.2km side each of them :D Your map has awesome speed of scan. Best that I ever see earlier.
ps. Ssory, eng not my native

[phatal]

also some times I getting error
[CRITICAL] 07.09.16 14:08:12 Database corrupted or locked
I take a look at uniq table for each thread and merging them after, but it will slow down whole scan process =\

[phatal]

One more question, why you don't implement cover_region to utils? It can be very useful when we sending geojson or list of points for region and getting scan list.

And may be you have poosibility to add mode for example pokemon/pokestops scan? (stops scans at 1km radius)

[tcmaps]

uhm, you can already make a config.json what it will load from
but cover_region can not cover geojson in the python s2sphere

[phatal]

I do already the same thing but just txt files with line by line data, problem that I'm want run much instances at one time. And when I do it (call bbotstrap through js) i getting issue with dividing by zero (of cause I can create own db for each instance but that not good imho. I running about 100+ instances with 400+ accs at same time. ) By the way, may be your calculate which number of accs at 1sq km (for example) optimal? when I trying looks like that my 50 instances with 1 accs by instance scan 50 points with 900m radius for 1:17, and when I using 4 accs per instances it take just 10sec less (about 1:06). Mostly time using login into accs. Problem with same IP at much accs I decide using multiple thread instances of TOR (1 by thread)

Aw... I just see https://github.com/mapbox/node-s2 this, and there have a possibility to covering array of points (polygon) with s2 cells. It base on original "C" library

And what about scan radius? As I remember in original pogomap it depend on "pulse" variable at search.py? Or it's just fake? Just when I asking at forum's about crapping pokestops and gyms all as one saying to me that enough 1km scan instead of 70m for porekemons. Where I can configure it?

[tcmaps]

you're running the instances from the same dir/db?
don't do that. sqlite can't handle concurrent writes,
that's why it's crashing with 'db locked'
just put accounts in accounts.txt