Fix runtime exception for unknown location_types at the same lat/lon

http://code.google.com/p/googletransitdatafeed/issues/detail?id=113

[clancy.childs, 2008-12-15 19:36:09]

Please review this change that fixes a RuntimeException when an unknown
location_type is found in the stop hierarchy validation. This now allows this to
pass, as the Stop.SetAndCheckAttr function will warn about the invalid type.
I am fixing this because even if the validator finds a bad value, it should just
report it and keep processing, rather than completely failing.

[Tom, 2008-12-15 20:48:43]

http://codereview.appspot.com/11051/diff/1/2
File python/test/testtransitfeed.py (right):

http://codereview.appspot.com/11051/diff/1/2#newcode1262
Line 1262: "BADLOCATION,Airport Bad Location,36.868446,-116.784582,2,\n"
Shorter is better: give STATION a bad location_type and remove BADLOCATION,
BULLFROG, STAGECOACH

http://codereview.appspot.com/11051/diff/1/3
File python/transitfeed.py (right):

http://codereview.appspot.com/11051/diff/1/3#newcode3055
Line 3055: elif (stop.location_type not in [0,1] or
Instead of having a negative match and an extra branch how about making the else
on 3060

elif stop.location_type in (0, 1) and other_stop.location_type  in (0, 1):

so that the conditions are consistent. location_type not in (0,1) will not match
any conditon and drop off the bottom.

Stop._CheckAndSetAttr use a tuple (0, 1) instead of list for valid
location_type. I don't think we want to factor out the list of valid types
because we might want a valid type that isn't checked for proximity.

[clancy.childs, 2008-12-16 00:20:10]

I have updated with Tom's suggestions.

http://codereview.appspot.com/11051/diff/1/2
File python/test/testtransitfeed.py (right):

http://codereview.appspot.com/11051/diff/1/2#newcode1262
Line 1262: "BADLOCATION,Airport Bad Location,36.868446,-116.784582,2,\n"
On 2008/12/15 20:48:44, Tom wrote:
> Shorter is better: give STATION a bad location_type and remove BADLOCATION,
> BULLFROG, STAGECOACH

I did this, but didn't remove BULLFROG and STAGECOACH as they are used by the
stop_times and trips that the class inherits and to remove them, would require
testing for more errors or re-doing the trips and stop_times for this test.

http://codereview.appspot.com/11051/diff/1/3
File python/transitfeed.py (right):

http://codereview.appspot.com/11051/diff/1/3#newcode3055
Line 3055: elif (stop.location_type not in [0,1] or
On 2008/12/15 20:48:44, Tom wrote:
> Instead of having a negative match and an extra branch how about making the
else
> on 3060
> 
> elif stop.location_type in (0, 1) and other_stop.location_type  in (0, 1):
> 
> so that the conditions are consistent. location_type not in (0,1) will not
match
> any conditon and drop off the bottom.
> 
> Stop._CheckAndSetAttr use a tuple (0, 1) instead of list for valid
> location_type. I don't think we want to factor out the list of valid types
> because we might want a valid type that isn't checked for proximity.

Done.

[Tom, 2008-12-16 01:42:06]

One nit pick and this is good to submit.

http://codereview.appspot.com/11051/diff/1/2
File python/test/testtransitfeed.py (right):

http://codereview.appspot.com/11051/diff/1/2#newcode1262
Line 1262: "BADLOCATION,Airport Bad Location,36.868446,-116.784582,2,\n"
On 2008/12/16 00:20:10, clancy.childs wrote:
> On 2008/12/15 20:48:44, Tom wrote:
> > Shorter is better: give STATION a bad location_type and remove BADLOCATION,
> > BULLFROG, STAGECOACH
> 
> I did this, but didn't remove BULLFROG and STAGECOACH as they are used by the
> stop_times and trips that the class inherits and to remove them, would require
> testing for more errors or re-doing the trips and stop_times for this test.

Shucks. I didn't look at testBadLocationType closely and just saw that it had
less stop rows. A MemoryZipTestCase with less preinitialized objects would be
handy but doesn't need to be in this change.

http://codereview.appspot.com/11051/diff/203/9
File python/transitfeed.py (right):

http://codereview.appspot.com/11051/diff/203/9#newcode3060
Line 3060: elif stop.location_type in (0, 1) and other_stop.location_type  in
(0, 1):
please wrap to 80 cols