Catch error if trying to load zero bytes resume file

Currently, when loading a resume file, it assumes that dill.load() will work successfully. This requires that the file is a valid pickle file, even if it an empty file (as generated by bilby_pipe for file transfer reasons). However, if would be useful if this could just be a generic zero bytes dummy file as output by, e.g., touch.

This MR adds try...except calls around dill.load() to catch cases of zero bytes files, or checks for check point files being 0 bytes in size.

bilby/core/sampler/ptemcee.py

@@ -415,7 +415,11 @@ class Ptemcee(MCMCSampler):
             # This is a very ugly hack to support numpy>=1.24
             ptemcee.sampler.np.float = float
 
-        if os.path.isfile(self.resume_file) and self.resume is True:
+        if (
+            os.path.isfile(self.resume_file)
+            and os.path.getsize(self.resume_file)
+            and self.resume is True
+        ):
             import dill
 
             logger.info(f"Resume data {self.resume_file} found")
@@ -513,7 +517,7 @@ class Ptemcee(MCMCSampler):
         logger.info("Starting to sample")
 
         while True:
-            for (pos0, log_posterior, log_likelihood) in sampler.sample(
+            for pos0, log_posterior, log_likelihood in sampler.sample(
                 self.pos0,
                 storechain=False,
                 iterations=self.convergence_inputs.niterations_per_check,