Add PTC data

.github/workflows/black_check.yml

@@ -23,4 +23,4 @@ jobs:
         pip install black==24.10.0
     - name: Run Black
       run: |
-        black --check .
+        black --check --target-version py39 .

.github/workflows/isort_check.yml

@@ -1,4 +1,4 @@
-name: Black
+name: Isort
 
 on:
   push:
@@ -24,4 +24,4 @@ jobs:
     - name: Isort check
       run: |
         isort --check --diff factgenie || echo "We know it fails ATM"
-        isort --check --diff tests 
+        isort --check --diff tests

factgenie/config/resources.yml

@@ -122,3 +122,21 @@ xsum-debug:
   source: https://github.com/EdinburghNLP/XSum
   splits:
   - test
+propaganda-techniques:
+  name: propaganda-techniques
+  class: propaganda_techniques.PropagandaTechniques
+  annotations:
+      - PTC
+  description: 'The Propaganda Techniques Corpus (PTC) is a corpus of propagandistic techniques at annotated by 6 pro-annotators in spans.
+    The corpus includes 451 articles (350k tokens) from 48 news outlets. 
+    The corpus accompanied paper [Fine-Grained Analysis of Propaganda in News Article](https://aclanthology.org/D19-1565/).
+    The labels are:
+Loaded Language, Name Calling&Labeling, Repetition, Exaggeration&Minimization, Doubt, Appeal to fear-prejudice, Flag-Waving, Causal Oversimplification, Slogans, Appeal to Authority, Black-and-White Fallacy, Thought-terminating Cliches, Whataboutism, Reductio ad Hitlerum, Red Herring, Bandwagon, Obfuscation&Intentional&Vagueness&Confusion, Straw Men'
+  data-link: https://owncloud.cesnet.cz/index.php/s/e5OKV8XF9LaQATj/download
+  source: https://web.archive.org/web/20210922205158/https://propaganda.qcri.org/data/fine_grained_propaganda_code.zip
+  outputs:
+    - identical 
+  splits:
+    - train
+    - dev
+    - test

factgenie/data/datasets_TEMPLATE.yml

@@ -1,6 +1,7 @@
 # ---
 
 # If you want to start adding local datasets manually, rename this template file to `datasets.yml` and modify the example record.
+# See the factgenie/data/resources.yml or the /manage/ page after you started factgenie server in order to see out-of-box supported datasets. 
 
 # A dataset record looks like the following:
 example-dataset-id: