Juan Luis Cano Rodríguez

Upstream response was "this is expected". Anyway, I would say this is also our fault in the `GeoVector.__eq__` method, which only expects another `GeoVector`: https://github.com/satellogic/telluric/blob/7c4d75e8f2f5407b5e65436d21339db78ecc61ef/telluric/vectors.py#L431-L438

This is similar to what attrs does: http://www.attrs.org/en/stable/why.html#hand-written-classes

As simple as: ```diff diff --git telluric/plotting.py telluric/plotting.py index f48cc10..c90f35b 100644 --- telluric/plotting.py +++ telluric/plotting.py @@ -135,7 +135,7 @@ def plot(feature, mp=None, style_function=None, **map_kwargs): else: if mp is None: center...

...15 is an arbitrary value that, for the case of our default basemap (stamen terrain), does not always contain data in all regions.

Thanks a lot for the good bug report @jmarintur! Could you please give the output of `crs.raster` for completeness?

Thanks @drnextgis! I have a couple of questions: 1. I understand that crop and reproject are not meant to be commutative. Reprojecting a big image can be extremely expensive, especially...

@drnextgis I agree. However, perhaps we could connect this to #43 in some way and produce a proper error when the conversion fails? Or, instead of an error, do a...

The histograms part is in a sorry state at the moment. Let's talk to @slava-kerner and @arielze to do some cleanup there. They are also related to the `stretch_histogram` functionality...

> we should solve the issue on `_raster_opener` and implement it as a context manager Something like this? ```python class RasterOpener: def __init__(self, filename, *args, *, **kwargs): try: self.raster =...

> On the other hand I think my solution would enforce using as contextmenager and cause all usages as regular function to fail. Yep, I tried that and it makes...