Disable tqdm in non-interactive mode (stanford-crfm#1422)

Author: yifanmai

scripts/fact_completion/create_benchmark.py

@@ -76,7 +76,7 @@ def main():
     # Select head QIDS to include in dataset
     selected = []
     all_qids = set()
-    for p in tqdm(prop_map):
+    for p in tqdm(prop_map, disable=None):
         np.random.seed(0)
         head_qids = list(prop_map[p].keys())
         idxs = np.random.choice(len(head_qids), min(1000, len(head_qids)), replace=False)

scripts/fact_completion/fetch_triples_and_aliases.py

@@ -80,7 +80,7 @@ def main() -> None:
     table_files.extend(get_batch_files(processed_folder / args.entity_values_folder))
 
     filtered_triples = []
-    for filepath in tqdm(table_files):
+    for filepath in tqdm(table_files, disable=None):
         filtered_triples.extend(rel_filtering_func(seed_relations, filepath))
     print(f"Extracted {len(filtered_triples)} triples for seed relations.")
 
@@ -96,7 +96,7 @@ def main() -> None:
 
     table_files = get_batch_files(processed_folder / args.aliases_folder)
     filtered_aliases = []
-    for filepath in tqdm(table_files):
+    for filepath in tqdm(table_files, disable=None):
         filtered_aliases.extend(alias_filtering_func(qids, filepath))
     print(f"Extracted {len(filtered_aliases)} aliases.")
 

scripts/fact_completion/filter_triples.py

@@ -73,7 +73,7 @@ def main() -> None:
     print("Loading Wikipedia table.")
     wikipedia_qids = set()
     table_files = get_batch_files(processed_folder / args.wikipedia_links_folder)
-    for f in tqdm(table_files):
+    for f in tqdm(table_files, disable=None):
         with open(f) as in_file:
             for line in in_file:
                 item = json.loads(line)
@@ -84,7 +84,7 @@ def main() -> None:
     triples_file = benchmark_folder / "triples.jsonl"
     print(f"Loading triples from {triples_file}. Filtering based on alias quality...")
     filtered_triples = []
-    for triple in tqdm(jsonl_generator(str(triples_file)), total=215288006):
+    for triple in tqdm(jsonl_generator(str(triples_file)), total=215288006, disable=None):
         q1 = triple["qid"]
         q2 = triple["value"]
         if bad_alias(qid2alias[q1]) or bad_alias(qid2alias[q2]):

src/helm/benchmark/presentation/summarize.py

@@ -291,7 +291,7 @@ def read_runs(self):
         # run_suite_path can contain subdirectories that are not runs (e.g. eval_cache, groups)
         # so filter them out.
         run_dir_names = sorted([p for p in os.listdir(self.run_suite_path) if p != "eval_cache" and p != "groups"])
-        for run_dir_name in tqdm(run_dir_names):
+        for run_dir_name in tqdm(run_dir_names, disable=None):
             run_spec_path: str = os.path.join(self.run_suite_path, run_dir_name, "run_spec.json")
             stats_path: str = os.path.join(self.run_suite_path, run_dir_name, "stats.json")
             if not os.path.exists(run_spec_path) or not os.path.exists(stats_path):

src/helm/benchmark/runner.py

@@ -96,7 +96,7 @@ def __init__(
         ensure_directory_exists(self.eval_cache_path)
 
     def run_all(self, run_specs: List[RunSpec]):
-        for run_spec in tqdm(run_specs):
+        for run_spec in tqdm(run_specs, disable=None):
             try:
                 with htrack_block(f"Running {run_spec.name}"):
                     self.run_one(run_spec)

src/helm/benchmark/scenarios/copyright_scenario.py

@@ -72,7 +72,7 @@ def get_instances(self) -> List[Instance]:
 
         # Read all the instances
         instances: List[Instance] = []
-        for prefix, prefix_to_end in tqdm.tqdm(data["data"].items(), desc="load instances"):
+        for prefix, prefix_to_end in tqdm.tqdm(data["data"].items(), desc="load instances", disable=None):
             instances.append(
                 Instance(
                     input=Input(text=prefix),

src/helm/common/general.py

@@ -195,13 +195,13 @@ def parallel_map(
     with htrack_block(f"Parallelizing computation on {len(items)} items over {parallelism} {units}"):
         results: List
         if parallelism == 1:
-            results = list(tqdm(map(process, items), total=len(items)))
+            results = list(tqdm(map(process, items), total=len(items), disable=None))
         elif multiprocessing:
             with ProcessPoolExecutor(max_workers=parallelism) as executor:
-                results = list(tqdm(executor.map(process, items), total=len(items)))
+                results = list(tqdm(executor.map(process, items), total=len(items), disable=None))
         else:
             with ThreadPoolExecutor(max_workers=parallelism) as executor:
-                results = list(tqdm(executor.map(process, items), total=len(items)))
+                results = list(tqdm(executor.map(process, items), total=len(items), disable=None))
     return results