Update filename handling in InferableTypeStage for filespec (#198)

* Update filename handling in InferableTypeStage for filespec

* Validate 'path' in filename for InferableTypeStage

Raise ValueError if 'path' is missing in filename dict

* Enhance tests for pdal.Reader and pdal.Writer types

Added assertions to validate reader and writer types for filespec format.

* support and test full filespec definitions

* conditionalize FileSpec support to PDAL 2.9+

* check headers for FileSpec

---------

Co-authored-by: Howard Butler <hobu.inc@gmail.com>

Author: mrtgenet

src/pdal/pipeline.py

@@ -33,6 +33,13 @@
 LogLevelFromPDAL = {v: k for k, v in LogLevelToPDAL.items()}
 
 
+HaveFileSpecSupport = False
+if libpdalpython.getInfo().major == 2 and \
+        libpdalpython.getInfo().minor >= 9 or \
+        libpdalpython.getInfo().major > 2:
+            HaveFileSpecSupport = True
+
+
 class Pipeline(libpdalpython.Pipeline):
     def __init__(
         self,
@@ -217,7 +224,25 @@ def __or__(self, other: Union[Stage, Pipeline]) -> Pipeline:
 class InferableTypeStage(Stage):
     def __init__(self, filename: Optional[str] = None, **options: Any):
         if filename:
-            options["filename"] = filename
+            if isinstance(filename, dict):
+                if "path" not in filename:
+                    raise ValueError(f"'path' is missing in the provided filespec: {filename}")
+                if HaveFileSpecSupport:
+                    options["filename"] = filename
+                else:
+                    # log that we can't pass FileSpec to PDAL
+                    # because the library version is too old
+                    msg = "PDAL library version is too old for FileSpec support. " \
+                          "Defaulting to using filename only"
+                    logging.info(msg)
+                    options["filename"] = filename["path"]
+
+            else:
+                if HaveFileSpecSupport:
+                    filespec = {'path':str(filename)}
+                    options["filename"] = filespec
+                else:
+                    options["filename"] = filename
         super().__init__(**options)
 
     @property
@@ -226,7 +251,18 @@ def type(self) -> str:
             return super().type
         except KeyError:
             filename = self._options.get("filename")
-            return str(self._infer_type(filename) if filename else "")
+            if isinstance(filename, dict):
+                if "path" not in filename:
+                    raise ValueError(f"'path' is missing in the provided filespec: {filename}")
+                if HaveFileSpecSupport:
+                    path = filename.get('path')
+                else:
+                    path = filename
+
+            else:
+                path = str(filename)
+
+            return str(self._infer_type(path) if filename else "")
 
     _infer_type = staticmethod(lambda filename: "")
 

test/test_pipeline.py

@@ -3,7 +3,6 @@
 import os
 import sys
 
-from typing import List
 from itertools import product
 import numpy as np
 import pytest
@@ -284,6 +283,20 @@ def test_infer_stage_type(self):
         assert pdal.Writer("foo.xxx").type == ""
         assert pdal.Reader().type == ""
         assert pdal.Writer().type == ""
+        assert pdal.Reader({"path": "foo.las"}).type == "readers.las"
+        assert pdal.Writer({"path": "foo.las"}).type == "writers.las"
+        assert pdal.Reader({"path": "foo.xxx"}).type == ""
+        assert pdal.Writer({"path": "foo.xxx"}).type == ""
+        assert pdal.Reader({}).type == ""
+        assert pdal.Writer({}).type == ""
+
+    def test_filespec(self):
+        """Can transit filespecs"""
+        spec = {'path':'junk.las', 'headers':{'header1':'header_1', 'header2':'header_2'}, 'query':{'query1':'query_1', 'query2':'query_2'}}
+        assert pdal.Reader(spec).type == "readers.las"
+        assert pdal.Reader(spec).options['filename']['path'] == "junk.las"
+        assert pdal.Reader(spec).options['filename']['headers']['header2'] == "header_2"
+        assert pdal.Writer("foo.las").type == "writers.las"
 
     def test_streamable(self):
         """Can we distinguish streamable from non-streamable stages and pipeline"""