From c799f2a4c776cc86f04075b4018abbd0cac65f6c Mon Sep 17 00:00:00 2001
From: Lukas Heinrich <lukas.heinrich@gmail.com>
Date: Mon, 19 Jul 2021 15:41:16 +0200
Subject: [PATCH] commit

---
 setup.py                                 |  1 +
 src/recastatlas/config.py                | 48 +++++++++++++++++-------
 src/recastatlas/subcommands/catalogue.py |  7 ++--
 3 files changed, 39 insertions(+), 17 deletions(-)

diff --git a/setup.py b/setup.py
index 35a0bc3..d1c8a1d 100644
--- a/setup.py
+++ b/setup.py
@@ -20,6 +20,7 @@
   include_package_data = True,
   install_requires = [
     'click',
+    'jsonschema',
     'pyyaml',
     'yadage-schemas==0.10.6',
   ],
diff --git a/src/recastatlas/config.py b/src/recastatlas/config.py
index d3fe910..cc6182d 100644
--- a/src/recastatlas/config.py
+++ b/src/recastatlas/config.py
@@ -3,6 +3,7 @@
 import pkg_resources
 import glob
 import logging
+import jsonschema
 
 log = logging.getLogger(__name__)
 
@@ -76,23 +77,42 @@ def catalogue(self):
         files = list(set(files))
         log.debug(files)
         for f in files:
-            d = yaml.safe_load(open(f))
-            if not validate_catalogue_entry(d):
-                continue
             log.debug(f'loading catalogue file {f}')
-            name = d.pop("name")
-            if not "toplevel" in d["spec"]:
-                d["spec"]["toplevel"] = os.path.realpath(
-                    os.path.join(os.path.dirname(f), "specs")
-                )
-            cfg[name] = d
+            entry = yaml.safe_load(open(f))
+            process_entry(cfg,f,entry)
         return cfg
 
-
 config = Config()
 
-def validate_catalogue_entry(entry):
-    for x in ['name','metadata','spec']:
-        if not x in entry:
-            return False
+def process_entry(cfg,fname,entry,tags = None):
+    if (entry is not None) and is_entry_file(entry):
+        name = entry.pop("name")
+        if not "toplevel" in entry["spec"]:
+            entry["spec"]["toplevel"] = os.path.realpath(
+                os.path.join(os.path.dirname(fname), "specs")
+            )
+            entry['metadata'].setdefault('tags',[]).extend(tags or [])
+        cfg[name] = entry
+    elif (entry is not None) and is_index_file(entry):
+        for index_f in entry['recast_catalogue_entries']:
+            index_f = os.path.join(os.path.dirname(fname),index_f)
+            log.debug(f'loading catalogue file {index_f}')
+            index_entry = yaml.safe_load(open(index_f))
+            process_entry(cfg,index_f,index_entry,entry['tags'])
+            
+def is_entry_file(entry):
+    schema = jsonschema.Draft7Validator({'required': ['name','metadata','spec']})    
+    try:
+        schema.validate(entry)
+    except jsonschema.exceptions.ValidationError:
+        return False
+    return True
+
+def is_index_file(entry):
+    schema = jsonschema.Draft7Validator({'required': ['name','tags','recast_catalogue_entries']})    
+    try:
+        schema.validate(entry)
+    except jsonschema.exceptions.ValidationError:
+        return False
     return True
+
diff --git a/src/recastatlas/subcommands/catalogue.py b/src/recastatlas/subcommands/catalogue.py
index aafb407..45b080a 100644
--- a/src/recastatlas/subcommands/catalogue.py
+++ b/src/recastatlas/subcommands/catalogue.py
@@ -13,7 +13,7 @@
 
 
 log = logging.getLogger(__name__)
-default_meta = {"author": "unknown", "short_description": "no description"}
+default_meta = {"author": "unknown", "short_description": "no description", "tags": []}
 
 
 @click.group(help="The RECAST Analysis Catalogue")
@@ -87,8 +87,8 @@ def rm(path):
 
 @catalogue.command()
 def ls():
-    fmt = "{0:35}{1:60}{2:20}"
-    click.secho(fmt.format("NAME", "DESCRIPTION", "EXAMPLES"))
+    fmt = "{0:35}{1:60}{2:20}{3:20}"
+    click.secho(fmt.format("NAME", "DESCRIPTION", "EXAMPLES", "TAGS"))
 
     for k, v in sorted(config.catalogue.items(), key=lambda x: x[0]):
         click.secho(
@@ -96,6 +96,7 @@ def ls():
                 k,
                 v.get("metadata", default_meta)["short_description"],
                 ",".join(list(v.get("example_inputs", {}).keys())),
+                ",".join(v.get("metadata", default_meta).get('tags',[]))
             )
         )