plot_results.py

# %%
import re
from pathlib import Path
from typing import List
import matplotlib.pyplot as plt


# %%
def extract_field(log_path: str, field: str) -> List[str]:
    """
    Extract values like:  field=value  from each line in the log.
    Works for fields such as:
      best_mrr, best_epoch, model, epochs, batch_size, lr, criterion, R@1, R@5, R@10, checkpoint, etc.
    Returns values as strings (convert to float/int if you want).
    """
    text = Path(log_path).read_text(encoding="utf-8", errors="ignore").splitlines()

    # Escape field for regex (important for things like "R@1")
    field_escaped = re.escape(field)

    # Capture until next separator: space OR " |" OR end of line
    # Handles "field=0.123", "field=mixed", "field=output/file.pt"
    pattern = re.compile(rf"\b{field_escaped}=(?P<val>.*?)(?=\s\|\s|\s+$|$)")

    out = []
    for line in text:
        m = pattern.search(line)
        if m:
            out.append(float(m.group("val").strip()))
    return out


# %%
log_file = "output/log_train.log"
field = "best_mrr"  # change to whatever you want

values = extract_field(log_file, field)
values
# %%
x = list(range(len(values)))
vals = sorted(values)
plt.scatter(x, vals)
plt.scatter([x[0]], [vals[0]], color="red", zorder=5, label="reference")
plt.plot(sorted(values), linestyle="--", label="MRR improvements")
plt.xlabel("Training Sorted")
plt.ylabel("MRR results")
plt.legend()
plt.show()
# %%
for i in [1, 5, 10]:
    field = f"R@{i}"  # change to whatever you want

    values = extract_field(log_file, field)
    vals = sorted(values)
    plt.scatter(x, vals)
    plt.scatter(
        [x[0]], [vals[0]], color="red", zorder=5, label="reference" if i == 1 else None
    )
    plt.plot(sorted(values), linestyle="--", label=field)
plt.xlabel("Training Sorted")
plt.ylabel("Recall rate")
plt.axhline(y=1.0, label="y=1")
plt.legend()
plt.show()
# %%