fix: perform sanity checks for datasets (close #27)

CHANGELOG

@@ -2,10 +2,12 @@
  - feat: implement download manager
  - feat: rename first tab to "Find Data" and allow searching for
    private datasets
+ - fix: perform sanity checks for datasets (#27)
  - enh: allow to download all resources of a collection (#13)
  - enh: implement search for packages the user has access to
  - enh: implement DBInterrogator functions for generating a private
    user db (partly #5)
+ - setup: bump dclab from 0.35.3 to 0.35.5
  - ref: simplify `dbmodel` submodule (just use Interrogator)
  - ref: expand dedicated `api` submodule (moved dunctions from
    `upload.dataset`)

dcoraid/upload/job.py

@@ -294,9 +294,17 @@ def task_compress_resources(self):
         self.set_state("compress")
         for ii, path in enumerate(list(self.paths)):
             if path.suffix in [".rtdc", ".dc"]:  # do we have an .rtdc file?
-                # check for compression
                 with IntegrityChecker(path) as ic:
+                    # check for compression
                     cdata = ic.check_compression()[0].data
+                    # perform a sanity check
+                    insane = [c.msg for c in ic.sanity_check()]
+                    if insane:
+                        # The user is responsible of cleaning up the mess.
+                        # We just make sure no dirty data gets uploaded to
+                        # DCOR.
+                        raise IOError(f"Sanity check failed for {path}!\n"
+                                      + "\n".join(insane))
                 if cdata["uncompressed"]:  # (partially) not compressed?
                     res_dir = self.cache_dir / "{}".format(ii)
                     res_dir.mkdir(exist_ok=True, parents=True)

setup.py

@@ -25,7 +25,7 @@
     description=description,
     long_description=open('README.rst').read() if exists('README.rst') else '',
     install_requires=["appdirs",  # cache directory for data compression
-                      "dclab[dcor]==0.35.3",  # pin for triage
+                      "dclab[dcor]==0.35.5",  # pin for triage
                       "pyqt5",
                       "requests>=2.13",
                       "requests_toolbelt",  # multipart uploads with progress