core: support '' for explicitly set empty path set

doc/MODULES.org

@@ -46,7 +46,9 @@ Some explanations:
   - =/a/path/to/directory/=, so the module will consume all files from this directory
   - a list of files/directories (it will be flattened)
   - a [[https://docs.python.org/3/library/glob.html?highlight=glob#glob.glob][glob]] string, so you can be flexible about the format of your data on disk (e.g. if you want to keep it compressed)
-  - empty sequence (e.g. ~export_path = ()~), this is useful for modules that merge multiple data sources (for example, =my.twitter=)
+  - empty string (e.g. ~export_path = ''~), this will prevent the module from consuming any data
+
+    This can be useful for modules that merge multiple data sources (for example, =my.twitter= or =my.github=)
 
   Typically, such variable will be passed to =get_files= to actually extract the list of real files to use. You can see usage examples [[https://github.com/karlicoss/HPI/blob/master/tests/get_files.py][here]].
 

doc/SETUP.org

@@ -474,8 +474,7 @@ Since you have two different sources of raw data, you need to specify two bits o
 : class twitter_archive:
 :     export_path = '/backups/twitter-archives/*.zip'
 
-Note that you can also just use =my.twitter.archive= or =my.twitter.twint= directly, or set either of paths to 'empty path': =()=
-# TODO empty string?
+Note that you can also just use =my.twitter.archive= or =my.twitter.twint= directly, or set either of paths to empty string: =''=
 # (TODO mypy-safe?)
 
 # #addingmodifying-modules

my/core/common.py

@@ -125,11 +125,16 @@ def get_files(pp: Paths, glob: str=DEFAULT_GLOB, sort: bool=True) -> Tuple[Path,
     Tuple as return type is a bit friendlier for hashing/caching, so hopefully makes sense
     """
     # TODO FIXME mm, some wrapper to assert iterator isn't empty?
-    sources: List[Path] = []
-    if isinstance(pp, (str, Path)):
-        sources.append(Path(pp))
+    sources: List[Path]
+    if isinstance(pp, Path):
+        sources = [pp]
+    elif isinstance(pp, str):
+        if pp == '':
+            # special case -- makes sense for optional data sources, etc
+            return () # early return to prevent warnings etc
+        sources = [Path(pp)]
     else:
-        sources.extend(map(Path, pp))
+        sources = [Path(p) for p in pp]
 
     def caller() -> str:
         import traceback

tests/get_files.py

@@ -102,6 +102,9 @@ def test_no_files():
     '''
     Test for empty matches. They work, but should result in warning
     '''
+    assert get_files('')         == ()
+
+    # todo test these for warnings?
     assert get_files([])         == ()
     assert get_files('bad*glob') == ()