X-Git-Url: https://git.arvados.org/rnaseq-cwl-training.git/blobdiff_plain/9c447b8e6ca87637b4ada4a6361c56d06cd8f0f8..a56ac2da1d6b7eb472212ae9c1d2c7057acf0f95:/bin/util.py

diff --git a/bin/util.py b/bin/util.py
index 23245be..92fd50d 100644
--- a/bin/util.py
+++ b/bin/util.py
@@ -1,24 +1,53 @@
 import sys
+import os
 import json
-import yaml
 from subprocess import Popen, PIPE
 
+# Import this way to produce a more useful error message.
+try:
+    import yaml
+except ImportError:
+    print('Unable to import YAML module: please install PyYAML', file=sys.stderr)
+    sys.exit(1)
+
+
+# Things an image file's name can end with.
+IMAGE_FILE_SUFFIX = {
+    '.gif',
+    '.jpg',
+    '.png',
+    '.svg'
+}
+
+# Files that shouldn't be present.
+UNWANTED_FILES = [
+    '.nojekyll'
+]
+
+# Marker to show that an expected value hasn't been provided.
+# (Can't use 'None' because that might be a legitimate value.)
+REPORTER_NOT_SET = []
 
 class Reporter(object):
     """Collect and report errors."""
 
-    def __init__(self, args):
+    def __init__(self):
         """Constructor."""
 
         super(Reporter, self).__init__()
         self.messages = []
 
 
-    def check_field(self, filename, name, values, key, expected):
+    def check_field(self, filename, name, values, key, expected=REPORTER_NOT_SET):
         """Check that a dictionary has an expected value."""
 
         if key not in values:
             self.add(filename, '{0} does not contain {1}', name, key)
+        elif expected is REPORTER_NOT_SET:
+            pass
+        elif type(expected) in (tuple, set, list):
+            if values[key] not in expected:
+                self.add(filename, '{0} {1} value {2} is not in {3}', name, key, values[key], expected)
         elif values[key] != expected:
             self.add(filename, '{0} {1} is {2} not {3}', name, key, values[key], expected)
 
@@ -51,27 +80,24 @@ class Reporter(object):
 
         if not self.messages:
             return
-        for m in self.messages:
+        for m in sorted(self.messages):
             print(m, file=stream)
 
 
 def read_markdown(parser, path):
-    """Get YAML and AST for Markdown file, returning {'metadata':yaml, 'text': text, 'doc':doc}."""
+    """
+    Get YAML and AST for Markdown file, returning
+    {'metadata':yaml, 'metadata_len':N, 'text':text, 'lines':[(i, line, len)], 'doc':doc}.
+    """
 
     # Split and extract YAML (if present).
-    metadata = None
-    metadata_len = None
     with open(path, 'r') as reader:
         body = reader.read()
-    pieces = body.split('---', 2)
-    if len(pieces) == 3:
-        try:
-            metadata = yaml.load(pieces[1])
-        except yaml.YAMLError as e:
-            print('Unable to parse YAML header in {0}:\n{1}'.format(path, e))
-            sys.exit(1)
-        metadata_len = pieces[1].count('\n')
-        body = pieces[2]
+    metadata_raw, metadata_yaml, body = split_metadata(path, body)
+
+    # Split into lines.
+    metadata_len = 0 if metadata_raw is None else metadata_raw.count('\n')
+    lines = [(metadata_len+i+1, line, len(line)) for (i, line) in enumerate(body.split('\n'))]
 
     # Parse Markdown.
     cmd = 'ruby {0}'.format(parser)
@@ -80,8 +106,66 @@ def read_markdown(parser, path):
     doc = json.loads(stdout_data)
 
     return {
-        'metadata': metadata,
+        'metadata': metadata_yaml,
         'metadata_len': metadata_len,
         'text': body,
+        'lines': lines,
         'doc': doc
     }
+
+
+def split_metadata(path, text):
+    """
+    Get raw (text) metadata, metadata as YAML, and rest of body.
+    If no metadata, return (None, None, body).
+    """
+
+    metadata_raw = None
+    metadata_yaml = None
+    metadata_len = None
+
+    pieces = text.split('---', 2)
+    if len(pieces) == 3:
+        metadata_raw = pieces[1]
+        text = pieces[2]
+        try:
+            metadata_yaml = yaml.load(metadata_raw)
+        except yaml.YAMLError as e:
+            print('Unable to parse YAML header in {0}:\n{1}'.format(path, e), file=sys.stderr)
+            sys.exit(1)
+
+    return metadata_raw, metadata_yaml, text
+
+
+def load_yaml(filename):
+    """
+    Wrapper around YAML loading so that 'import yaml' is only needed
+    in one file.
+    """
+
+    try:
+        with open(filename, 'r') as reader:
+            return yaml.load(reader)
+    except (yaml.YAMLError, FileNotFoundError) as e:
+        print('Unable to load YAML file {0}:\n{1}'.format(filename, e), file=sys.stderr)
+        sys.exit(1)
+
+
+def check_unwanted_files(dir_path, reporter):
+    """
+    Check that unwanted files are not present.
+    """
+
+    for filename in UNWANTED_FILES:
+        path = os.path.join(dir_path, filename)
+        reporter.check(not os.path.exists(path),
+                       path,
+                       "Unwanted file found")
+
+
+def require(condition, message):
+    """Fail if condition not met."""
+
+    if not condition:
+        print(message, file=sys.stderr)
+        sys.exit(1)