Merge branch 'master' of github.com:TREMA-UNH/trec-car-tools

.travis.yml

@@ -1,9 +1,18 @@
 language: python
+python:
+  - "3.5"
 before_install:
   - sudo apt-get -qq update
-  - sudo apt-get install -y maven2 python3-pip
+  - sudo apt-get install -y maven
+python:
+  - "3.4"
+  - "3.5"
+  - "3.6"
+  - "3.7"
+install:
+  - pip install -r python3/requirements.txt
 script:
-  - pip3 install --user python3/
+  - pip install python3/
   - pushd trec-car-tools-example; mvn install; popd
 
   - curl http://trec-car.cs.unh.edu/datareleases/v2.0/test200.v2.0.tar.xz | tar -xJ

python3/test.py

@@ -1,4 +1,4 @@
-#!/usr/bin/python3
+#!/usr/bin/env python3
 
 from trec_car.read_data import *
 import argparse

python3/trec_car/read_data.py

@@ -129,7 +129,11 @@ def from_cbor(cbor):
         elif typetag == 1: return CategoryPage()
         elif typetag == 2: return DisambiguationPage()
         elif typetag == 3:
-            targetPage = cbor[1].decode('ascii')
+            target = cbor[1]
+            if type(target) == list: # TODO this is almost certainly wrong
+                targetPage = target[1]
+            else:
+                targetPage = target.decode('ascii')
             return RedirectPage(targetPage)
         else:
             print("Deserialisation error for PageType cbor="+cbor)
@@ -228,7 +232,7 @@ def default():
     def __str__(self):
         redirStr = ("" if self.redirectNames is None else (" redirected = "+", ".join([name for name in self.redirectNames])))
         disamStr = ("" if self.disambiguationNames is None else (" disambiguated = "+", ".join([name for name in self.disambiguationNames])))
-        catStr = ("" if self.redirectNames is None else (" categories = "+", ".join([name for name in self.categoryNames])))
+        catStr = ("" if self.redirectNames is None else (" categories = "+", ".join([name for name in (self.categoryNames or [])])))
         inlinkStr = ("" if self.inlinkIds is None else (" inlinks = "+", ".join([name for name in self.inlinkIds])))
         # inlinkAnchorStr = str (self.inlinkAnchors)
         inlinkAnchorStr = ("" if self.inlinkAnchors is None else \