Create figure in current directory #273
Conversation
| @@ -295,7 +295,8 @@ def run_likelihood() -> None: | |||
| [a.legend(fontsize="small") for a in ax] | |||
|
|
|||
| fig.suptitle("PT Cls, including IA, galaxy bias, magnification") | |||
| fig.savefig("plots/pt_cls.png", facecolor="white", dpi=300) | |||
| with open("pt_cls.png", "w+b") as png_file: | |||
| fig.savefig(png_file, facecolor="white", dpi=300) | |||
There was a problem hiding this comment.
Why the context manager instead of using the file name directly in savefig?
There was a problem hiding this comment.
Habit. I prefer to make it obvious the resource is “cleaned up” (in this case, the file closed) rather than having a reader need to know whether the routine being called does the right thing.
There was a problem hiding this comment.
If it's just here that fine but generally I'd argue having a context manager for savefig is overkill and likely to confuse users (what happens if you forget the "b"?). Might as well throw in a os.makedirs("plots", exists_ok=True) to deal with the subdir in that case.
| @@ -295,7 +295,8 @@ def run_likelihood() -> None: | |||
| [a.legend(fontsize="small") for a in ax] | |||
|
|
|||
| fig.suptitle("PT Cls, including IA, galaxy bias, magnification") | |||
| fig.savefig("plots/pt_cls.png", facecolor="white", dpi=300) | |||
| with open("pt_cls.png", "w+b") as png_file: | |||
| fig.savefig(png_file, facecolor="white", dpi=300) | |||
There was a problem hiding this comment.
If it's just here that fine but generally I'd argue having a context manager for savefig is overkill and likely to confuse users (what happens if you forget the "b"?). Might as well throw in a os.makedirs("plots", exists_ok=True) to deal with the subdir in that case.
vitenti
deleted the
272-example-des_y1_3x2pt_ptpy-crashes-when-run-as-a-script
branch
This fixes the issue by creating the figure in the current directory, which was simpler than creating a subdirectory.