[GSProcessing] Fix parsing of file paths that start with "./" (#1091)

*Issue #, if available:*

*Description of changes:*

* Some GConstruct or GSProcessing config files from customers could have
file paths that start with `./` which works when input is local but
break for S3. We fix this by removing the `./` prefix if it exists when
parsing the input files.
* Add change to config we use for testing to verify the change works
(although we'd probably need a more S3-specific test)

By submitting this pull request, I confirm that you can use, modify,
copy, and redistribute this contribution, under the terms of your
choice.

graphstorm-processing/graphstorm_processing/config/data_config_base.py

@@ -15,7 +15,7 @@
 """
 
 from dataclasses import dataclass
-from typing import Sequence, Optional
+from typing import Optional
 
 from graphstorm_processing.constants import SUPPORTED_FILE_TYPES
 
@@ -27,7 +27,7 @@ class DataStorageConfig:
     """
 
     format: str
-    files: Sequence[str]
+    files: list[str]
     separator: Optional[str] = None
 
     def __post_init__(self):
@@ -39,3 +39,7 @@ def __post_init__(self):
                     raise ValueError(
                         f"File paths need to be relative (not starting with '/'), got : {file}"
                     )
+
+        for idx, file in enumerate(self.files):
+            if file.startswith("./"):
+                self.files[idx] = file[2:]

graphstorm-processing/graphstorm_processing/graph_loaders/dist_heterogeneous_loader.py

@@ -939,7 +939,7 @@ def process_node_data(self, node_configs: Sequence[NodeConfig]) -> Dict:
         self.graph_info["ntype_to_label_masks"] = defaultdict(list)
         for node_config in node_configs:
             files = node_config.files
-            file_paths = [f"{self.input_prefix}/{f}" for f in files]
+            file_paths = [os.path.join(self.input_prefix, f) for f in files]
 
             node_type = node_config.ntype
             node_col = node_config.node_col

graphstorm-processing/tests/resources/small_heterogeneous_graph/gsprocessing-config.json

@@ -17,7 +17,7 @@
                 "data": {
                     "format": "csv",
                     "files": [
-                        "nodes/movie.csv"
+                        "./nodes/movie.csv"
                     ],
                     "separator": ","
                 },
@@ -104,7 +104,7 @@
                 "data": {
                     "format": "csv",
                     "files": [
-                        "edges/movie-included_in-genre.csv"
+                        "./edges/movie-included_in-genre.csv"
                     ],
                     "separator": ","
                 },