Sebastian Gehrmann
commited on
Commit
•
1bfed91
1
Parent(s):
fe1f9e0
- Taskmaster.py +34 -33
Taskmaster.py
CHANGED
@@ -12,7 +12,7 @@
|
|
12 |
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
13 |
# See the License for the specific language governing permissions and
|
14 |
# limitations under the License.
|
15 |
-
"""
|
16 |
|
17 |
|
18 |
import csv
|
@@ -36,12 +36,18 @@ _CITATION = """\
|
|
36 |
# TODO: Add description of the dataset here
|
37 |
# You can copy an official description
|
38 |
_DESCRIPTION = """\
|
39 |
-
The Taskmaster-3 (aka TicketTalk) dataset consists of 23,789 movie ticketing dialogs
|
40 |
-
|
|
|
|
|
|
|
|
|
41 |
"""
|
42 |
|
43 |
# TODO: Add a link to an official homepage for the dataset here
|
44 |
-
_HOMEPAGE =
|
|
|
|
|
45 |
|
46 |
# TODO: Add the licence for the dataset here if you can find it
|
47 |
_LICENSE = "CC BY 4.0"
|
@@ -73,7 +79,6 @@ class Taskmaster(datasets.GeneratorBasedBuilder):
|
|
73 |
|
74 |
VERSION = datasets.Version("3.0.0")
|
75 |
|
76 |
-
|
77 |
# If you need to make complex sub-parts in the datasets with configurable options
|
78 |
# You can create your own builder configuration class to store attribute, inheriting from datasets.BuilderConfig
|
79 |
# BUILDER_CONFIG_CLASS = MyBuilderConfig
|
@@ -88,7 +93,7 @@ class Taskmaster(datasets.GeneratorBasedBuilder):
|
|
88 |
# datasets.TaskmasterConfig(name="test", version=VERSION, description="test set"),
|
89 |
# ]
|
90 |
|
91 |
-
#DEFAULT_CONFIG_NAME = "TaskmasterConfig" # It's not mandatory to have a default configuration. Just use one if it makes sense.
|
92 |
|
93 |
def _info(self):
|
94 |
# TODO: This method specifies the datasets.DatasetInfo object which contains informations and typings for the dataset
|
@@ -134,23 +139,22 @@ class Taskmaster(datasets.GeneratorBasedBuilder):
|
|
134 |
# dl_manager is a datasets.download.DownloadManager that can be used to download and extract URLs
|
135 |
# It can accept any type or nested list/dict and will give back the same structure with the url replaced with path to local files.
|
136 |
# By default the archives will be extracted and a path to a cached folder where they are extracted is returned instead of the archive
|
137 |
-
#_URLs = { "train": "train.csv", "test": "test.csv", "validation": "dev.csv", }
|
138 |
-
my_urls = {
|
|
|
|
|
|
|
|
|
139 |
data_dir = dl_manager.download_and_extract(my_urls)
|
140 |
-
challenge_sets = [
|
141 |
-
# ("challenge_%s_nov_%s" % (split,lvl),"%s-%s_nv2_%s.jsonl" % (split,self.config.name,lvl)) \
|
142 |
-
# for split in ["train","valid","test"] for lvl in ["low","mid","high"]
|
143 |
-
("challenge_%s_nov_%s" % (split,lvl),"%s-%s_nv2_%s.csv" % (split,self.config.name,lvl)) \
|
144 |
-
for split in ["train","valid","test"] for lvl in ["low","mid","high"]
|
145 |
-
]
|
146 |
-
# + ...
|
147 |
|
148 |
return [
|
149 |
datasets.SplitGenerator(
|
150 |
name=datasets.Split.TRAIN,
|
151 |
# These kwargs will be passed to _generate_examples
|
152 |
gen_kwargs={
|
153 |
-
"filepath": os.path.join(
|
|
|
|
|
154 |
"split": "train",
|
155 |
},
|
156 |
),
|
@@ -158,38 +162,35 @@ class Taskmaster(datasets.GeneratorBasedBuilder):
|
|
158 |
name=datasets.Split.TEST,
|
159 |
# These kwargs will be passed to _generate_examples
|
160 |
gen_kwargs={
|
161 |
-
"filepath": os.path.join(
|
162 |
-
|
|
|
|
|
163 |
},
|
164 |
),
|
165 |
datasets.SplitGenerator(
|
166 |
name=datasets.Split.VALIDATION,
|
167 |
# These kwargs will be passed to _generate_examples
|
168 |
gen_kwargs={
|
169 |
-
"filepath": os.path.join(
|
|
|
|
|
170 |
"split": "dev",
|
171 |
},
|
172 |
),
|
173 |
-
]
|
174 |
-
datasets.SplitGenerator(
|
175 |
-
name=challenge_split,
|
176 |
-
gen_kwargs={
|
177 |
-
"filepath": os.path.join(data_dir[challenge_split], filename),
|
178 |
-
"split": challenge_split,
|
179 |
-
},
|
180 |
-
)
|
181 |
-
for challenge_split, filename in challenge_sets
|
182 |
-
]
|
183 |
|
184 |
def _generate_examples(
|
185 |
-
self,
|
|
|
|
|
186 |
):
|
187 |
-
"""
|
188 |
# This method handles input defined in _split_generators to yield (key, example) tuples from the dataset.
|
189 |
# The `key` is here for legacy reason (tfds) and is not important in itself.
|
190 |
|
191 |
with open(filepath, encoding="utf-8") as f:
|
192 |
for row in f:
|
193 |
data = csv.loads(row)
|
194 |
-
data["gem_id"] = "GEM-TASKMASTER-%s-%d" % (split,data["id"]+1)
|
195 |
-
yield data["id"],data
|
|
|
12 |
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
13 |
# See the License for the specific language governing permissions and
|
14 |
# limitations under the License.
|
15 |
+
"""A movie ticketing dialog dataset with 23,789 annotated conversations.."""
|
16 |
|
17 |
|
18 |
import csv
|
|
|
36 |
# TODO: Add description of the dataset here
|
37 |
# You can copy an official description
|
38 |
_DESCRIPTION = """\
|
39 |
+
The Taskmaster-3 (aka TicketTalk) dataset consists of 23,789 movie ticketing dialogs
|
40 |
+
(located in Taskmaster/TM-3-2020/data/). By "movie ticketing" we mean conversations
|
41 |
+
where the customer's goal is to purchase tickets after deciding on theater, time,
|
42 |
+
movie name, number of tickets, and date, or opt out of the transaction.
|
43 |
+
The columns are gem_id, 0, 1 for serial numbering, 2 for the text dialog and id
|
44 |
+
for the default id by the authors.
|
45 |
"""
|
46 |
|
47 |
# TODO: Add a link to an official homepage for the dataset here
|
48 |
+
_HOMEPAGE = (
|
49 |
+
"https://github.com/google-research-datasets/Taskmaster/tree/master/TM-3-2020"
|
50 |
+
)
|
51 |
|
52 |
# TODO: Add the licence for the dataset here if you can find it
|
53 |
_LICENSE = "CC BY 4.0"
|
|
|
79 |
|
80 |
VERSION = datasets.Version("3.0.0")
|
81 |
|
|
|
82 |
# If you need to make complex sub-parts in the datasets with configurable options
|
83 |
# You can create your own builder configuration class to store attribute, inheriting from datasets.BuilderConfig
|
84 |
# BUILDER_CONFIG_CLASS = MyBuilderConfig
|
|
|
93 |
# datasets.TaskmasterConfig(name="test", version=VERSION, description="test set"),
|
94 |
# ]
|
95 |
|
96 |
+
# DEFAULT_CONFIG_NAME = "TaskmasterConfig" # It's not mandatory to have a default configuration. Just use one if it makes sense.
|
97 |
|
98 |
def _info(self):
|
99 |
# TODO: This method specifies the datasets.DatasetInfo object which contains informations and typings for the dataset
|
|
|
139 |
# dl_manager is a datasets.download.DownloadManager that can be used to download and extract URLs
|
140 |
# It can accept any type or nested list/dict and will give back the same structure with the url replaced with path to local files.
|
141 |
# By default the archives will be extracted and a path to a cached folder where they are extracted is returned instead of the archive
|
142 |
+
# _URLs = { "train": "train.csv", "test": "test.csv", "validation": "dev.csv", }
|
143 |
+
my_urls = {
|
144 |
+
"train": "train.csv",
|
145 |
+
"test": "test.csv",
|
146 |
+
"validation": "dev.csv",
|
147 |
+
} # _URLs[self.config.name]
|
148 |
data_dir = dl_manager.download_and_extract(my_urls)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
149 |
|
150 |
return [
|
151 |
datasets.SplitGenerator(
|
152 |
name=datasets.Split.TRAIN,
|
153 |
# These kwargs will be passed to _generate_examples
|
154 |
gen_kwargs={
|
155 |
+
"filepath": os.path.join(
|
156 |
+
data_dir["train"], "train-%s.csv"
|
157 |
+
), # % (self.config.name)),
|
158 |
"split": "train",
|
159 |
},
|
160 |
),
|
|
|
162 |
name=datasets.Split.TEST,
|
163 |
# These kwargs will be passed to _generate_examples
|
164 |
gen_kwargs={
|
165 |
+
"filepath": os.path.join(
|
166 |
+
data_dir["test"], "test-%s.csv"
|
167 |
+
), # % (self.config.name)),
|
168 |
+
"split": "test",
|
169 |
},
|
170 |
),
|
171 |
datasets.SplitGenerator(
|
172 |
name=datasets.Split.VALIDATION,
|
173 |
# These kwargs will be passed to _generate_examples
|
174 |
gen_kwargs={
|
175 |
+
"filepath": os.path.join(
|
176 |
+
data_dir["validation"], "valid-%s.csv"
|
177 |
+
), # % (self.config.name)),
|
178 |
"split": "dev",
|
179 |
},
|
180 |
),
|
181 |
+
]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
182 |
|
183 |
def _generate_examples(
|
184 |
+
self,
|
185 |
+
filepath,
|
186 |
+
split, # method parameters are unpacked from `gen_kwargs` as given in `_split_generators`
|
187 |
):
|
188 |
+
"""Yields examples as (key, example) tuples."""
|
189 |
# This method handles input defined in _split_generators to yield (key, example) tuples from the dataset.
|
190 |
# The `key` is here for legacy reason (tfds) and is not important in itself.
|
191 |
|
192 |
with open(filepath, encoding="utf-8") as f:
|
193 |
for row in f:
|
194 |
data = csv.loads(row)
|
195 |
+
data["gem_id"] = "GEM-TASKMASTER-%s-%d" % (split, data["id"] + 1)
|
196 |
+
yield data["id"], data
|