add functionality required for this test load_collection & collect

add functionality required for this test load_collection & collect

https://github.com/apache/incubator-wayang/blob/1e0f9e8166225176fe3022de5fbcce3dbcba96b9/python/src/pywy/tests/test_train_linear_svc.py#L24

```python

# limitations under the License.
#

import pytest

from pywy.dataquanta import WayangContext
from pywy.platforms.java import JavaPlugin
from pywy.platforms.spark import SparkPlugin

# TODO: add functionality required for this test load_collection & collect
@pytest.mark.skip(reason="no way of currently testing this, since we are missing implementations for load_collection & collect")
def test_train_and_predict():
ctx = WayangContext().register({JavaPlugin, SparkPlugin})

features = ctx.load_collection([[0.0, 1.0], [1.0, 0.0], [1.0, 1.0], [0.0, 0.0]])
labels = ctx.load_collection([1.0, 1.0, 0.0, 0.0])

model = features.train_linear_svc(labels, max_iter=10, reg_param=0.1)
predictions = model.predict(features)

result = predictions.collect()

print("Predictions:", result)

assert len(result) is 4, f"Expected len(result) to be 4, but got: {len(result)}"
for pred in result:
assert pred in [0.0, 1.0], f"Expected prediction to be in [0.0, 0.1], but got: {pred}"

```

6bd117a221c0fa18b566075684463e7058390aca

Start with python/src/pywy/tests/test_train_linear_svc.py at the skipped test and trace the missing load_collection and collect entry points. Read the surrounding pywy dataquanta and registered JavaPlugin and SparkPlugin APIs before deciding where these operations belong. Done means the test can be unskipped and its predictions collection satisfies the stated assertions.

Written by the indexing model from the issue text.