Parsing unknown wkt parsing now works (though only very basic detection method), and better error handling in tester

pycrs/loader.py

@@ -16,7 +16,7 @@ def from_url(url, format=None):
 
     Arguments:
     - *url*: The url where the crs string is to be read from. 
-    - *format*: Which format to parse the crs string as. One of "ogcwkt", "esriwkt", or "proj4", but also several others...
+    - *format*: Which format to parse the crs string as. One of "ogc wkt", "esri wkt", or "proj4".
 
     Returns:
     - CRS object.

pycrs/parser.py

@@ -95,14 +95,30 @@ def from_esri_wkt(string, strict=False):
     # parse arguments into components
     # use args to create crs
     return _from_wkt(string, "esri", strict)
-        
+
+def from_unknown_wkt(string, strict=False):
+    """
+    Given an unknown wkt string, detect if uses ogc or esri flavor, and parse the crs accordingly.
+
+    Arguments:
+    - *string*: The unknown WKT representation as a string.
+    - *strict* (optional): When True, the parser is strict about names having to match
+        exactly with upper and lowercases. Default is not strict (False).
+
+    Returns:
+    - CRS object.
+    """
+    # parse arguments into components
+    # use args to create crs
+    return _from_wkt(string, None, strict)
+
 def _from_wkt(string, wkttype, strict=False):
     """
     Internal method for parsing wkt, with minor differences depending on ogc or esri style.
 
     Arguments:
     - *string*: The OGC or ESRI WKT representation as a string.
-    - *wkttype*: How to parse the WKT string, as either 'ogc' or 'esri'.
+    - *wkttype* (optional): How to parse the WKT string, as either 'ogc', 'esri', or None. If None, tries to autodetect the wkt type before parsing (default). 
     - *strict* (optional): When True, the parser is strict about names having to match
         exactly with upper and lowercases. Default is not strict (False).
 
@@ -114,8 +130,8 @@ def _from_wkt(string, wkttype, strict=False):
     # - Maybe verify elem arg name
 
     # make sure valid wkttype
-    wkttype = wkttype.lower()
-    assert wkttype in ("ogc","esri")
+    if wkttype: wkttype = wkttype.lower()
+    assert wkttype in ("ogc","esri",None)
     
     # remove newlines and multi spaces
     string = " ".join(string.split())
@@ -212,7 +228,7 @@ def _from_wkt(string, wkttype, strict=False):
         consumed = _clean_value(consumed)
         items.append(consumed)
         return items
-
+    
     # load into nested tuples and arglists
     crstuples = []
     chars = (char for char in string)
@@ -222,6 +238,24 @@ def _from_wkt(string, wkttype, strict=False):
         crstuples.append((header, content))
         char = next(chars, None)
 
+    # autodetect wkttype if not specified
+    if not wkttype:
+        topheader,topcontent = crstuples[0]
+        if topheader == "PROJCS":
+            geogcsheader,geogcscontent = topcontent[1]
+        elif topheader == "GEOGCS":
+            geogcsheader,geogcscontent = topheader,topcontent
+
+        # datum elem should be second under geogcs
+        datumheader, datumcontent = geogcscontent[1]
+        datumname = datumcontent[0].upper().strip('"')
+        
+        # esri wkt datums all use "D_" before the datum name
+        if datumname.startswith("D_"):
+            wkttype = "esri"
+        else:
+            wkttype = "ogc"
+
     # parse into actual crs objects
     def _parse_top(header, content):
         "procedure for parsing the toplevel crs element and all its children"
@@ -396,12 +430,6 @@ def _from_wkt(string, wkttype, strict=False):
     # use args to create crs
     return crs
 
-##def from_unknown_wkt(string, strict=False):
-##    # detect if ogc wkt or esri wkt
-##    # TIPS: esri wkt datums all use "D_" before the datum name
-##    # then load with appropriate function
-##    pass
-
 def from_proj4(string, strict=False):
     """
     Parse crs as proj4 formatted string and return the resulting crs object.
@@ -717,23 +745,23 @@ def from_unknown_text(text, strict=False):
     Detect crs string format and parse into crs object with appropriate function.
 
     Arguments:
-    - *string*: The crs text representation of unknown type.
+    - *string*: The crs text representation of unknown type. 
     - *strict* (optional): When True, the parser is strict about names having to match
         exactly with upper and lowercases. Default is not strict (False).
 
     Returns:
     - CRS object.
     """
-    
-    if string.startswith("urn:"):
-        from_ogc_urn(string, strict)
 
-    elif string.startswith("+"):
+    if string.startswith("+"):
         from_proj4(string, strict)
 
     elif string.startswith(("PROJCS[","GEOGCS[")):
         from_unknown_wkt(string, strict)
 
+    #elif string.startswith("urn:"):
+    #    from_ogc_urn(string, strict)
+
     elif string.startswith("EPSG:"):
         from_epsg_code(string.split(":")[1])
 

tester.py

@@ -1,5 +1,5 @@
 import pycrs
-
+import traceback
 
 
 
@@ -84,15 +84,15 @@ def sourcestrings(format):
 def testoutputs(crs):
     print("To:\n")
     try: result = crs.to_ogc_wkt()
-    except: result = "Fail"
+    except: result = traceback.format_exc()
     print("ogc_wkt: %s \n" % result)
 
     try: result = crs.to_esri_wkt()
-    except: result = "Fail"
+    except: result = traceback.format_exc()
     print("esri_wkt: %s \n" % result)
           
     try: result = crs.to_proj4()
-    except: result = "Fail"
+    except: result = traceback.format_exc()
     print("proj4: %s \n" % result)
 
 
@@ -129,8 +129,8 @@ for wkt in sourcestrings("ogcwkt"):
         # test outputs
         testoutputs(crs)
         
-    except Exception as err:
-        print(err)   
+    except:
+        print(traceback.format_exc()+"\n")   
 
 #render_world(crs)