SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1911.07249 · 2019

The Proper Care and Feeding of CAMELS: How Limited Training Data Affects Streamflow Prediction

arXiv · PDF · Open in the Atlas

Code that ran

13 of 14 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 12 have no contract verdict; 1 failure of our instrument; 1 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/1911.07249.json · badge: /api/ran/1911.07249.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
gauchm/ealstm_regional_modeling canonical (classifier label)
classifier basis: census
13 ran · 1 no recorded run
FunctionStatusWhere it lives
calc_alpha_nse Ran
ran on a synthesized input
gauchm/ealstm_regional_modeling/papercode/metrics.py
code served (permissive licence) · get_code("1991f7b4961d51bd")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned float64, ndim 0, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Create sample observation and simulation data
obs_data = np.array([1.2, 2.3, 3.1, 4.5, 2.8, 1.9])
sim_data = np.array([1.1, 2.5, 3.0, 4.3, 2.9, 2.1])

ARGS = {
    "obs": obs_data,
    "sim": sim_data
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
calc_beta_nse Ran
ran on a synthesized input
gauchm/ealstm_regional_modeling/papercode/metrics.py
code served (permissive licence) · get_code("12340ba2e594bd55")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned float64, ndim 0, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Create small numpy arrays for observations and simulations
# These represent discharge measurements over time
obs = np.array([1.5, 2.3, 1.8, 2.1, 1.9, 2.2])
sim = np.array([1.6, 2.4, 1.7, 2.0, 2.0, 2.3])

ARGS = {
    'obs': obs,
    'sim': sim
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
calc_nse Ran
ran on a synthesized input
gauchm/ealstm_regional_modeling/papercode/metrics.py
code served (permissive licence) · get_code("755c08582e820958")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned float64, ndim 0, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Create sample observation and simulation arrays
# Using different shapes to test the flatten operation
obs = np.array([[1.0, 2.0, 3.0], [4.0, 5.0, 6.0]])
sim = np.array([[1.1, 2.1, 2.9], [4.2, 4.8, 6.1]])

ARGS = {
    'obs': obs,
    'sim': sim
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
ecdf Ran
ran; our contract draft was wrong, not the code
gauchm/ealstm_regional_modeling/papercode/plotutils.py
code served (permissive licence) · get_code("56d870f7584b3428")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 2, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np
import torch

# Set random seed for reproducibility
torch.manual_seed(0)
np.random.seed(0)

# Create a simple numpy array with some data points
x = np.array([3.2, 1.5, 4.1, 2.8, 5.0, 1.2, 3.7, 2.1])

ARGS = {
    'x': x
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 2, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_clusters Ran
ran on a synthesized input
gauchm/ealstm_regional_modeling/papercode/clusterutils.py
code served (permissive licence) · get_code("040feb3b499435ae")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned dict. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np
import pandas as pd
from collections import defaultdict

# Create small sample data for testing
np.random.seed(0)

# Create lstm_features: Dict with basin IDs as keys, embedding arrays as values
# Each embedding should be shape (1, feature_dim) based on how it's used: emb.reshape(1, -1)
lstm_features = {
    '01013500': np.random.randn(1, 4),
    '01022500': np.random.randn(1, 4),
    '01031500': np.random.randn(1, 4),
}

# Create raw_features: pd.DataFrame with basin IDs as index and catchment attributes as columns
raw_features = pd.DataFrame(
    np.random.randn(3, 4),
    index=['01013500', '01022500', '01031500'],
    columns=['attr1', 'attr2', 'attr3', 'attr4']
)

# Create ks: List of cluster numbers
ks = [2, 3]

# Create basins: List of 8-digit USGS basin IDs
basins = ['01013500', '01022500', '01031500']

ARGS = {
    'lstm_features': lstm_features,
    'raw_features': raw_features,
    'ks': ks,
    'basins': basins,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, pandas, sklearn — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_label_2_color Ran
ran on a synthesized input
gauchm/ealstm_regional_modeling/papercode/clusterutils.py
code served (permissive licence) · get_code("1a855e3ea0dd9718")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned defaultdict. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

from collections import defaultdict

ARGS = {
    'lstm_clusters': {
        'basin_1': 0,
        'basin_2': 0,
        'basin_3': 1,
        'basin_4': 1,
        'basin_5': 2,
    },
    'raw_clusters': {
        'basin_1': 0,
        'basin_2': 1,
        'basin_3': 1,
        'basin_4': 2,
        'basin_5': 2,
    },
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type defaultdict — a defaultdict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a defaultdict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_run_dirs Ran
ran on a synthesized input
gauchm/ealstm_regional_modeling/papercode/evalutils.py
code served (permissive licence) · get_code("4c446d543ab69d8f")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import json
import tempfile
from pathlib import PosixPath

# Create a temporary directory structure for testing
temp_dir = tempfile.mkdtemp()
root_dir = PosixPath(temp_dir)

# Create subdirectories with cfg.json files for different model configurations
def create_run_folder(root, folder_name, cfg_dict):
    folder_path = root / folder_name
    folder_path.mkdir(exist_ok=True)
    cfg_file = folder_path / 'cfg.json'
    with open(cfg_file, 'w') as f:
        json.dump(cfg_dict, f)
    return folder_path

# Create sample folders with different configurations

# EALSTM with NSELoss
create_run_folder(
    root_dir,
    'run_ealstm_nse_1',
    {
        'concat_static': False,
        'no_static': False,
        'use_mse': False,
        'use_nse': True
    }
)

# LSTM with MSELoss
create_run_folder(
    root_dir,
    'run_lstm_mse_1',
    {
        'concat_static': True,
        'no_static': False,
        'use_mse': True,
        'use_nse': False
    }
)

# LSTM_no_static with NSELoss
create_run_folder(
    root_dir,
    'run_lstm_no_static_nse_1',
    {
        'concat_static': False,
        'no_static': True,
        'use_mse': False,
        'use_nse': True
    }
)

# XGBoost with MSELoss
create_run_folder(
    root_dir,
    'run_xgb_mse_1',
    {
        'concat_static': False,
        'no_static': False,
        'use_mse': True,
        'use_nse': False
    }
)

ARGS = {
    'root_dir': root_dir,
    'model': 'ealstm',
    'loss': 'NSELoss'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_run_dirs_gridEvaluation Ran
ran on a synthesized input
gauchm/ealstm_regional_modeling/papercode/evalutils.py
code served (permissive licence) · get_code("ecbb2deb80bb3950")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import json
import tempfile
from pathlib import PosixPath
from typing import List

# Create a temporary directory structure that simulates the grid evaluation setup
temp_root = tempfile.TemporaryDirectory()
root_dir = PosixPath(temp_root.name)

# Create sample run directories with proper naming convention
# Format appears to be: run_[model]_[other_fields]_[date]_[basin_subset]_[other]
sample_configs = [
    {
        'folder_name': 'run_ealstm_config_30092002_0_v1',
        'cfg': {
            'concat_static': False,
            'no_static': False,
            'use_mse': False
        }
    },
    {
        'folder_name': 'run_ealstm_config_30092005_0_v1',
        'cfg': {
            'concat_static': False,
            'no_static': False,
            'use_mse': False
        }
    },
    {
        'folder_name': 'run_ealstm_config_30092008_0_v1',
        'cfg': {
            'concat_static': False,
            'no_static': False,
            'use_mse': False
        }
    },
    {
        'folder_name': 'run_xgb_config_30092002_0_v1',
        'cfg': {
            'use_mse': False
        }
    },
    {
        'folder_name': 'run_xgb_config_30092005_0_v1',
        'cfg': {
            'use_mse': False
        }
    },
    {
        'folder_name': 'run_xgb_config_30092008_0_v1',
        'cfg': {
            'use_mse': False
        }
    },
]

# Create the folder structure
for config in sample_configs:
    folder_path = root_dir / config['folder_name']
    folder_path.mkdir(parents=True, exist_ok=True)
    cfg_file = folder_path / 'cfg.json'
    with open(cfg_file, 'w') as f:
        json.dump(config['cfg'], f)

# Create some invalid folders that should be skipped
invalid_folder = root_dir / 'run_ealstm_config_param_search_30092002_0_v1'
invalid_folder.mkdir(parents=True, exist_ok=True)
with open(invalid_folder / 'cfg.json', 'w') as f:
    json.dump({'concat_static': False, 'no_static': False, 'use_mse': False}, f)

ARGS = {
    'root_dir': root_dir,
    'model': 'ealstm',
    'basin_subset': 0,
    'training_years': 3
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_shape_collections Ran
ran on a synthesized input
gauchm/ealstm_regional_modeling/papercode/plotutils.py
code served (permissive licence) · get_code("f48dd78c80f393e6")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned PatchCollection. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Construct valid arguments for get_shape_collections
# The function expects a dict where each value has 'lons' and 'lats' keys
# containing coordinate arrays that can be transposed and used as polygon points

ARGS = {
    "data": {
        "region_1": {
            "lons": np.array([0.0, 1.0, 1.0, 0.0]),
            "lats": np.array([0.0, 0.0, 1.0, 1.0])
        },
        "region_2": {
            "lons": np.array([2.0, 3.0, 3.5, 2.5]),
            "lats": np.array([1.0, 1.0, 2.0, 2.0])
        }
    }
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
matplotlib, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type PatchCollection — a PatchCollection output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a PatchCollection output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_silhouette_scores Ran
ran on a synthesized input
gauchm/ealstm_regional_modeling/papercode/clusterutils.py
code served (permissive licence) · get_code("938143c5ecc5641d")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Create a small feature array for testing
# Shape: (n_samples, n_features) - using 20 samples with 5 features
features = np.random.RandomState(42).randn(20, 5)

# Define ARGS dict with the single required parameter
ARGS = {
    "features": features
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
load_attributes Ran
ran on a synthesized input
gauchm/ealstm_regional_modeling/papercode/datautils.py
code served (permissive licence) · get_code("0d729ca958fc0d3a")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned DataFrame. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import sqlite3
import tempfile
import pandas as pd
import os

# Create a temporary SQLite database with sample data
temp_db = tempfile.NamedTemporaryFile(delete=False, suffix='.db')
temp_db.close()
db_path = temp_db.name

# Create and populate the database
conn = sqlite3.connect(db_path)
cursor = conn.cursor()

# Create the basin_attributes table
cursor.execute('''
    CREATE TABLE basin_attributes (
        gauge_id INTEGER PRIMARY KEY,
        gauge_name TEXT,
        gauge_lat REAL,
        gauge_lon REAL,
        area_geospa_fabric REAL,
        geol_1st_class TEXT,
        glim_1st_class_frac REAL,
        geol_2nd_class TEXT,
        glim_2nd_class_frac REAL,
        dom_land_cover_frac REAL,
        dom_land_cover TEXT,
        high_prec_timing TEXT,
        low_prec_timing TEXT,
        huc TEXT,
        q_mean REAL,
        runoff_ratio REAL,
        stream_elas REAL,
        slope_fdc REAL,
        baseflow_index REAL,
        hfd_mean REAL,
        q5 REAL,
        q95 REAL,
        high_q_freq REAL,
        high_q_dur REAL,
        low_q_freq REAL,
        low_q_dur REAL,
        zero_q_freq REAL,
        geol_porostiy REAL,
        root_depth_50 REAL,
        root_depth_99 REAL,
        organic_frac REAL,
        water_frac REAL,
        other_frac REAL,
        feature_1 REAL,
        feature_2 REAL,
        feature_3 REAL
    )
''')

# Insert sample rows
cursor.execute('''
    INSERT INTO basin_attributes VALUES
    (12345678, 'gauge1', 40.5, -105.2, 100.0, 'class_a', 0.5, 'class_b', 0.3, 0.7, 'forest', 'spring', 'fall', '10130203', 50.0, 0.6, 1.2, 0.3, 0.4, 25.0, 10.0, 100.0, 0.5, 10.0, 0.2, 5.0, 0.1, 0.8, 1.5, 2.0, 0.3, 0.2, 0.1, 1.1, 2.2, 3.3),
    (87654321, 'gauge2', 41.3, -106.8, 150.0, 'class_c', 0.4, 'class_d', 0.4, 0.6, 'grassland', 'winter', 'summer', '10130204', 60.0, 0.7, 1.3, 0.4, 0.5, 30.0, 15.0, 120.0, 0.6, 12.0, 0.3, 6.0, 0.2, 0.85, 1.6, 2.1, 0.35, 0.25, 0.15, 1.2, 2.3, 3.4),
    (11111111, 'gauge3', 39.8, -104.5, 200.0, 'class_e', 0.6, 'class_f', 0.2, 0.8, 'urban', 'spring', 'summer', '10130205', 40.0, 0.5, 1.0, 0.2, 0.3, 20.0, 5.0, 90.0, 0.4, 8.0, 0.1, 4.0, 0.05, 0.75, 1.4, 1.9, 0.25, 0.15, 0.05, 1.0, 2.1, 3.2)
''')

conn.commit()
conn.close()

# Prepare arguments for the function
basins = [12345678, 87654321]
keep_features = ['feature_1', 'feature_2', 'feature_3']

ARGS = {
    'db_path': db_path,
    'basins': basins,
    'drop_lat_lon': True,
    'keep_features': keep_features
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
pandas — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DataFrame — a DataFrame output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DataFrame output the recorded ndim 2 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
normalize_features Ran
ran on a synthesized input
gauchm/ealstm_regional_modeling/papercode/datautils.py
code served (permissive licence) · get_code("c37a05f06f05d465")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 2, dtype family float, shape (2, 5) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Test case 1: normalize inputs (meteorological forcing data)
feature_inputs = np.array([
    [3.17563234, 372.01003929, 17.31934062, 3.97393362, 924.98004197],
    [5.0, 400.0, 20.0, 5.0, 1000.0]
], dtype=np.float32)

# Test case 2: normalize output (discharge values)
feature_output = np.array([[1.5], [2.0]], dtype=np.float32)

ARGS = {
    'feature': feature_inputs,
    'variable': 'inputs'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 2, dtype family float, shape (2, 5) — shape (2, 5) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
rescale_features Ran
ran on a synthesized input
gauchm/ealstm_regional_modeling/papercode/datautils.py
code served (permissive licence) · get_code("b62b4383d90e0390")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 2, dtype family float, shape (2, 5) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Create a small feature array for inputs (5 features as expected by SCALER)
feature_inputs = np.array([[0.1, 0.2, 0.3, 0.4, 0.5], 
                           [0.2, 0.3, 0.4, 0.5, 0.6]], dtype=np.float32)

# Create a small feature array for output (1 feature as expected by SCALER)
feature_output = np.array([[0.5], [0.7]], dtype=np.float32)

ARGS = {
    'feature': feature_inputs,
    'variable': 'inputs'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 2, dtype family float, shape (2, 5) — shape (2, 5) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_morris_gradient Not yet run
unverified (harvested, no recorded run)
gauchm/ealstm_regional_modeling/papercode/morris.py
code served (permissive licence) · get_code("7a44dc282997b5eb")
How it ran

no invocation recorded.

Also present in other papers' code

Of 14 samples lifted from this paper's linked repositories, 14 carry a function identity and 0 do not. Of the 14 distinct functions with an identity, 12 are also present in other papers whose code we hold (12 with at least one paper dated earlier, 0 the same month, 0 later) and 2 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

FunctionAlso present inEarliest-datedOthers named
calc_alpha_nse 1 other paper
1 dated earlier
1907.08456 (2019-07)
Towards Learning Universal, Regional, and Local Hydrological Behaviors via Machine-Learning Applied to Large-Sample Datasets
calc_beta_nse 1 other paper
1 dated earlier
1907.08456 (2019-07)
Towards Learning Universal, Regional, and Local Hydrological Behaviors via Machine-Learning Applied to Large-Sample Datasets
calc_nse 1 other paper
1 dated earlier
1907.08456 (2019-07)
Towards Learning Universal, Regional, and Local Hydrological Behaviors via Machine-Learning Applied to Large-Sample Datasets
ecdf 1 other paper
1 dated earlier
1907.08456 (2019-07)
Towards Learning Universal, Regional, and Local Hydrological Behaviors via Machine-Learning Applied to Large-Sample Datasets
get_clusters 1 other paper
1 dated earlier
1907.08456 (2019-07)
Towards Learning Universal, Regional, and Local Hydrological Behaviors via Machine-Learning Applied to Large-Sample Datasets
get_label_2_color 1 other paper
1 dated earlier
1907.08456 (2019-07)
Towards Learning Universal, Regional, and Local Hydrological Behaviors via Machine-Learning Applied to Large-Sample Datasets
get_morris_gradient 1 other paper
1 dated earlier
1907.08456 (2019-07)
Towards Learning Universal, Regional, and Local Hydrological Behaviors via Machine-Learning Applied to Large-Sample Datasets
get_shape_collections 1 other paper
1 dated earlier
1907.08456 (2019-07)
Towards Learning Universal, Regional, and Local Hydrological Behaviors via Machine-Learning Applied to Large-Sample Datasets
get_silhouette_scores 1 other paper
1 dated earlier
1907.08456 (2019-07)
Towards Learning Universal, Regional, and Local Hydrological Behaviors via Machine-Learning Applied to Large-Sample Datasets
load_attributes 1 other paper
1 dated earlier
1907.08456 (2019-07)
Towards Learning Universal, Regional, and Local Hydrological Behaviors via Machine-Learning Applied to Large-Sample Datasets
normalize_features 1 other paper
1 dated earlier
1907.08456 (2019-07)
Towards Learning Universal, Regional, and Local Hydrological Behaviors via Machine-Learning Applied to Large-Sample Datasets
rescale_features 1 other paper
1 dated earlier
1907.08456 (2019-07)
Towards Learning Universal, Regional, and Local Hydrological Behaviors via Machine-Learning Applied to Large-Sample Datasets

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Accurate streamflow prediction largely relies on historical meteorological records and streamflow measurements. For many regions, however, such data are only scarcely available. Facing this problem, many studies simply trained their machine learning models on the region's available data, leaving possible repercussions of this strategy unclear. In this study, we evaluate the sensitivity of tree- and LSTM-based models to limited training data, both in terms of geographic diversity and different time spans. We feed the models meteorological observations disseminated with the CAMELS dataset, and individually restrict the training period length, number of training basins, and input sequence length. We quantify how additional training data improve predictions and how many previous days of forcings we should feed the models to obtain best predictions for each training set size. Further, our findings show that tree- and LSTM-based models provide similarly accurate predictions on small datasets, while LSTMs are superior given more training data.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("1911.07249")
get_code_for_paper("1911.07249")
have("1911.07249")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/1911.07249.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/1911.07249.svg)](https://syntology.ai/paper/1911.07249)

Connect an agent — have() is free.