Use JSON output from verapdf

scripts/verapdf.py

@@ -9,6 +9,7 @@
 
 # USAGE: ./verapdf.py [$pdf_filepath|--process-all-test-pdf-files|--print-aggregated-report]
 
+import json
 import sys
 from subprocess import run, DEVNULL, PIPE
 
@@ -22,28 +23,30 @@ def analyze_pdf_file(pdf_filepath):
     command = [
         "verapdf/verapdf" + BAT_EXT,
         "--format",
-        "text",
-        "-v",
+        "json",
         pdf_filepath,
     ]
     # print(" ".join(command))
-    output = run(
-        command, stdout=PIPE, stderr=DEVNULL if HIDE_STDERR else None
-    ).stdout.decode()
-    # print(output)
-    return pdf_filepath, parse_output(output)
+    output = run(command, stdout=PIPE, stderr=DEVNULL if HIDE_STDERR else None)
+
+    if output.returncode != 0:
+        return pdf_filepath, {"failure": output.stdout.decode()}
+
+    return parse_output(output.stdout.decode())
 
 
 def parse_output(output):
     "Parse VeraPDF CLI output into a dict."
-    lines = output.splitlines()
-    try:
-        grave_line = next(line for line in lines if line.startswith("GRAVE:"))
-        return {"failure": grave_line}
-    except StopIteration:
-        # Skipping the first line
-        errors = [line[len("  FAIL ") :] for line in lines[1:]]
-        return {"errors": errors}
+    output_dict = json.loads(output)
+    assert len(output_dict["report"]["jobs"]) == 1
+    output_job = output_dict["report"]["jobs"][0]
+    file_path = output_job["itemDetails"]["name"]
+    rule_summaries = output_job["validationResult"]["details"]["ruleSummaries"]
+    errors = [
+        f"{rule_summary['clause']}-{rule_summary['testNumber']}"
+        for rule_summary in rule_summaries
+    ]
+    return file_path, {"errors": errors}
 
 
 if __name__ == "__main__":