Spaces:
Sleeping
Sleeping
Upload 10 files
Browse files- refresh_sources.py +24 -5
- sources.yaml +1 -14
refresh_sources.py
CHANGED
|
@@ -22,13 +22,32 @@ CH_URL = "http://127.0.0.1:8123"
|
|
| 22 |
|
| 23 |
def run_subprocess(cmd, cwd=None, env=None):
|
| 24 |
try:
|
| 25 |
-
|
|
|
|
|
|
|
| 26 |
cmd, cwd=cwd, env=env,
|
| 27 |
-
|
| 28 |
)
|
| 29 |
-
|
| 30 |
-
|
| 31 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 32 |
|
| 33 |
|
| 34 |
def execute_clickhouse_query(sql):
|
|
|
|
| 22 |
|
| 23 |
def run_subprocess(cmd, cwd=None, env=None):
|
| 24 |
try:
|
| 25 |
+
# We use Popen to stream stderr to the console for progress bars
|
| 26 |
+
# while still capturing stdout for return values if needed.
|
| 27 |
+
process = subprocess.Popen(
|
| 28 |
cmd, cwd=cwd, env=env,
|
| 29 |
+
stdout=subprocess.PIPE, stderr=subprocess.PIPE, text=True
|
| 30 |
)
|
| 31 |
+
|
| 32 |
+
# Stream stderr in real time
|
| 33 |
+
stderr_output = []
|
| 34 |
+
while True:
|
| 35 |
+
line = process.stderr.readline()
|
| 36 |
+
if not line and process.poll() is not None:
|
| 37 |
+
break
|
| 38 |
+
if line:
|
| 39 |
+
sys.stderr.write(line)
|
| 40 |
+
sys.stderr.flush()
|
| 41 |
+
stderr_output.append(line)
|
| 42 |
+
|
| 43 |
+
stdout, _ = process.communicate()
|
| 44 |
+
|
| 45 |
+
if process.returncode == 0:
|
| 46 |
+
return stdout.strip(), True
|
| 47 |
+
else:
|
| 48 |
+
return "".join(stderr_output).strip(), False
|
| 49 |
+
except Exception as e:
|
| 50 |
+
return str(e), False
|
| 51 |
|
| 52 |
|
| 53 |
def execute_clickhouse_query(sql):
|
sources.yaml
CHANGED
|
@@ -28,20 +28,7 @@ sources:
|
|
| 28 |
- view_name: hf_ohlc
|
| 29 |
file_glob: "**/*.parquet"
|
| 30 |
format: Parquet
|
| 31 |
-
description: "Full OHLCV data from HF dataset"
|
| 32 |
-
# columns: # explicit column list for the VIEW
|
| 33 |
-
# - timestamp
|
| 34 |
-
# - symbol
|
| 35 |
-
# - isin
|
| 36 |
-
# - series
|
| 37 |
-
# - open
|
| 38 |
-
# - high
|
| 39 |
-
# - low
|
| 40 |
-
# - close
|
| 41 |
-
# - volume
|
| 42 |
-
# - interval_minutes
|
| 43 |
-
# - segment
|
| 44 |
-
# - exchange
|
| 45 |
|
| 46 |
# ββ EXAMPLE: Adding a second data source ββββββββββββββββββββββββββββββ
|
| 47 |
# Just uncomment and edit:
|
|
|
|
| 28 |
- view_name: hf_ohlc
|
| 29 |
file_glob: "**/*.parquet"
|
| 30 |
format: Parquet
|
| 31 |
+
description: "Full OHLCV data from HF dataset, selects all 150+ columns automatically"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 32 |
|
| 33 |
# ββ EXAMPLE: Adding a second data source ββββββββββββββββββββββββββββββ
|
| 34 |
# Just uncomment and edit:
|