nextrec 0.4.27__tar.gz → 0.4.29__tar.gz

{nextrec-0.4.27 → nextrec-0.4.29}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: nextrec
-Version: 0.4.27
+Version: 0.4.29
 Summary: A comprehensive recommendation library with match, ranking, and multi-task learning models
 Project-URL: Homepage, https://github.com/zerolovesea/NextRec
 Project-URL: Repository, https://github.com/zerolovesea/NextRec
@@ -69,7 +69,7 @@ Description-Content-Type: text/markdown
 ![Python](https://img.shields.io/badge/Python-3.10+-blue.svg)
 ![PyTorch](https://img.shields.io/badge/PyTorch-1.10+-ee4c2c.svg)
 ![License](https://img.shields.io/badge/License-Apache%202.0-green.svg)
-![Version](https://img.shields.io/badge/Version-0.4.27-orange.svg)
+![Version](https://img.shields.io/badge/Version-0.4.29-orange.svg)
 [![Ask DeepWiki](https://deepwiki.com/badge.svg)](https://deepwiki.com/zerolovesea/NextRec)
 
 中文文档 | [English Version](README_en.md)
@@ -254,11 +254,11 @@ nextrec --mode=predict --predict_config=path/to/predict_config.yaml
 
 预测结果固定保存到 `{checkpoint_path}/predictions/{name}.{save_data_format}`。
 
-> 截止当前版本0.4.27，NextRec CLI支持单机训练，分布式训练相关功能尚在开发中。
+> 截止当前版本0.4.29，NextRec CLI支持单机训练，分布式训练相关功能尚在开发中。
 
 ## 兼容平台
 
-当前最新版本为0.4.27，所有模型和测试代码均已在以下平台通过验证，如果开发者在使用中遇到兼容问题，请在issue区提出错误报告及系统版本：
+当前最新版本为0.4.29，所有模型和测试代码均已在以下平台通过验证，如果开发者在使用中遇到兼容问题，请在issue区提出错误报告及系统版本：
 
 | 平台 | 配置 | 
 |------|------|

{nextrec-0.4.27 → nextrec-0.4.29}/README.md

@@ -8,7 +8,7 @@
 ![Python](https://img.shields.io/badge/Python-3.10+-blue.svg)
 ![PyTorch](https://img.shields.io/badge/PyTorch-1.10+-ee4c2c.svg)
 ![License](https://img.shields.io/badge/License-Apache%202.0-green.svg)
-![Version](https://img.shields.io/badge/Version-0.4.27-orange.svg)
+![Version](https://img.shields.io/badge/Version-0.4.29-orange.svg)
 [![Ask DeepWiki](https://deepwiki.com/badge.svg)](https://deepwiki.com/zerolovesea/NextRec)
 
 中文文档 | [English Version](README_en.md)
@@ -193,11 +193,11 @@ nextrec --mode=predict --predict_config=path/to/predict_config.yaml
 
 预测结果固定保存到 `{checkpoint_path}/predictions/{name}.{save_data_format}`。
 
-> 截止当前版本0.4.27，NextRec CLI支持单机训练，分布式训练相关功能尚在开发中。
+> 截止当前版本0.4.29，NextRec CLI支持单机训练，分布式训练相关功能尚在开发中。
 
 ## 兼容平台
 
-当前最新版本为0.4.27，所有模型和测试代码均已在以下平台通过验证，如果开发者在使用中遇到兼容问题，请在issue区提出错误报告及系统版本：
+当前最新版本为0.4.29，所有模型和测试代码均已在以下平台通过验证，如果开发者在使用中遇到兼容问题，请在issue区提出错误报告及系统版本：
 
 | 平台 | 配置 | 
 |------|------|

{nextrec-0.4.27 → nextrec-0.4.29}/README_en.md

@@ -8,7 +8,7 @@
 ![Python](https://img.shields.io/badge/Python-3.10+-blue.svg)
 ![PyTorch](https://img.shields.io/badge/PyTorch-1.10+-ee4c2c.svg)
 ![License](https://img.shields.io/badge/License-Apache%202.0-green.svg)
-![Version](https://img.shields.io/badge/Version-0.4.27-orange.svg)
+![Version](https://img.shields.io/badge/Version-0.4.29-orange.svg)
 [![Ask DeepWiki](https://deepwiki.com/badge.svg)](https://deepwiki.com/zerolovesea/NextRec)
 
 English | [中文文档](README.md)
@@ -196,11 +196,11 @@ nextrec --mode=predict --predict_config=path/to/predict_config.yaml
 
 Prediction outputs are saved under `{checkpoint_path}/predictions/{name}.{save_data_format}`.
 
-> As of version 0.4.27, NextRec CLI supports single-machine training; distributed training features are currently under development.
+> As of version 0.4.29, NextRec CLI supports single-machine training; distributed training features are currently under development.
 
 ## Platform Compatibility
 
-The current version is 0.4.27. All models and test code have been validated on the following platforms. If you encounter compatibility issues, please report them in the issue tracker with your system version:
+The current version is 0.4.29. All models and test code have been validated on the following platforms. If you encounter compatibility issues, please report them in the issue tracker with your system version:
 
 | Platform | Configuration | 
 |----------|---------------|

{nextrec-0.4.27 → nextrec-0.4.29}/docs/rtd/conf.py

@@ -11,7 +11,7 @@ sys.path.insert(0, str(PROJECT_ROOT / "nextrec"))
 project = "NextRec"
 copyright = "2025, Yang Zhou"
 author = "Yang Zhou"
-release = "0.4.27"
+release = "0.4.29"
 
 extensions = [
     "myst_parser",

nextrec-0.4.29/nextrec/__version__.py

@@ -0,0 +1 @@
+__version__ = "0.4.29"

{nextrec-0.4.27 → nextrec-0.4.29}/nextrec/basic/loggers.py

@@ -111,7 +111,7 @@ def setup_logger(session_id: str | os.PathLike | None = None):
     session = create_session(str(session_id) if session_id is not None else None)
     log_dir = session.logs_dir
     log_dir.mkdir(parents=True, exist_ok=True)
-    log_file = log_dir / "runs.log"
+    log_file = log_dir / "runs_log.txt"
 
     console_format = "%(message)s"
     file_format = "%(asctime)s - %(levelname)s - %(message)s"

{nextrec-0.4.27 → nextrec-0.4.29}/nextrec/data/batch_utils.py

@@ -5,13 +5,27 @@ Date: create on 03/12/2025
 Author: Yang Zhou, zyaztec@gmail.com
 """
 
-from typing import Any, Mapping
+from typing import Any, Mapping, Literal
 
 import numpy as np
 import torch
 
 
-def stack_section(batch: list[dict], section: str):
+def stack_section(batch: list[dict], section: Literal["features", "labels", "ids"]):
+    """   
+    input example:
+    batch = [
+        {"features": {"f1": tensor1, "f2": tensor2}, "labels": {"label": tensor3}},
+        {"features": {"f1": tensor4, "f2": tensor5}, "labels": {"label": tensor6}},
+        ...
+    ]
+    output example:
+    {
+        "f1": torch.stack([tensor1, tensor4], dim=0),
+        "f2": torch.stack([tensor2, tensor5], dim=0),
+    }
+    
+    """
     entries = [item.get(section) for item in batch if item.get(section) is not None]
     if not entries:
         return None
@@ -22,7 +36,13 @@ def stack_section(batch: list[dict], section: str):
             for item in batch
             if item.get(section) is not None and name in item[section]
         ]
-        merged[name] = torch.stack(tensors, dim=0)
+        tensor_sample = tensors[0]
+        if isinstance(tensor_sample, torch.Tensor):
+            merged[name] = torch.stack(tensors, dim=0)
+        elif isinstance(tensor_sample, np.ndarray):
+            merged[name] = np.stack(tensors, axis=0)
+        else:
+            merged[name] = tensors
     return merged
 
 

{nextrec-0.4.27 → nextrec-0.4.29}/nextrec/data/dataloader.py

@@ -2,7 +2,7 @@
 Dataloader definitions
 
 Date: create on 27/10/2025
-Checkpoint: edit on 24/12/2025
+Checkpoint: edit on 01/01/2026
 Author: Yang Zhou,zyaztec@gmail.com
 """
 
@@ -523,13 +523,8 @@ def build_tensors_from_data(
                 raise KeyError(
                     f"[RecDataLoader Error] ID column '{id_col}' not found in provided data."
                 )
-            try:
-                id_arr = np.asarray(column, dtype=np.int64)
-            except Exception as exc:
-                raise TypeError(
-                    f"[RecDataLoader Error] ID column '{id_col}' must contain numeric values. Received dtype={np.asarray(column).dtype}, error: {exc}"
-                ) from exc
-            id_tensors[id_col] = to_tensor(id_arr, dtype=torch.long)
+            # Normalize all id columns to strings for consistent downstream handling.
+            id_tensors[id_col] = np.asarray(column, dtype=str)
     if not feature_tensors:
         return None
     return {"features": feature_tensors, "labels": label_tensors, "ids": id_tensors}

{nextrec-0.4.27 → nextrec-0.4.29}/nextrec/models/multi_task/pepnet.py

@@ -61,9 +61,9 @@ from nextrec.utils.model import select_features
 from nextrec.utils.types import TaskTypeName
 
 
-class PPNetBlock(nn.Module):
+class PPNet(nn.Module):
     """
-    PEPNet block with per-layer gates conditioned on task context.
+    PPNet: per-task tower with layer-wise gates conditioned on task context.
     """
 
     def __init__(
@@ -274,18 +274,21 @@ class PEPNet(BaseModel):
         )
         task_dim = domain_dim + user_dim + item_dim
 
-        self.feature_gate = GateMLP(
+        # EPNet: shared feature-level gate (paper's EPNet).
+        self.epnet = GateMLP(
             input_dim=input_dim + domain_dim,
             hidden_dim=feature_gate_mlp_params["hidden_dim"],
             output_dim=input_dim,
             activation=feature_gate_mlp_params["activation"],
             dropout=feature_gate_mlp_params["dropout"],
             use_bn=feature_gate_mlp_params["use_bn"],
+            scale_factor=2.0,
         )
 
-        self.ppn_blocks = nn.ModuleList(
+        # PPNet: per-task gated towers (paper's PPNet).
+        self.ppnet_blocks = nn.ModuleList(
             [
-                PPNetBlock(
+                PPNet(
                     input_dim=input_dim,
                     output_dim=1,
                     gate_input_dim=input_dim + task_dim,
@@ -300,9 +303,9 @@ class PEPNet(BaseModel):
         self.prediction_layer = TaskHead(
             task_type=self.task, task_dims=[1] * self.nums_task
         )
-        self.grad_norm_shared_modules = ["embedding", "feature_gate"]
+        self.grad_norm_shared_modules = ["embedding", "epnet"]
         self.register_regularization_weights(
-            embedding_attr="embedding", include_modules=["feature_gate", "ppn_blocks"]
+            embedding_attr="embedding", include_modules=["epnet", "ppnet_blocks"]
         )
 
     def forward(self, x: dict[str, torch.Tensor]) -> torch.Tensor:
@@ -327,11 +330,11 @@ class PEPNet(BaseModel):
         task_sf_emb = torch.cat(task_parts, dim=-1)
 
         gate_input = torch.cat([dnn_input.detach(), domain_emb], dim=-1)
-        dnn_input = self.feature_gate(gate_input) * dnn_input
+        dnn_input = self.epnet(gate_input) * dnn_input
 
         task_logits = []
-        for block in self.ppn_blocks:
-            task_logits.append(block(o_ep=dnn_input, o_prior=task_sf_emb))
+        for block in self.ppnet_blocks:
+             task_logits.append(block(o_ep=dnn_input, o_prior=task_sf_emb))
 
         y = torch.cat(task_logits, dim=1)
         return self.prediction_layer(y)

{nextrec-0.4.27 → nextrec-0.4.29}/nextrec_cli_preset/NextRec-CLI.md

@@ -820,7 +820,7 @@ nextrec_logs/
     ├── features_config.pkl        # Feature configuration
     ├── model_epoch_1.pt        # Model checkpoint
     ├── model_epoch_2.pt
-    └── runs.log               # Training log
+    └── runs_log.txt               # Training log
 ```
 
 ### Q3: How to resume training?
@@ -829,7 +829,7 @@ The current CLI doesn't support resuming training. Please use the Python API.
 
 ### Q4: How to view training logs?
 
-Logs are saved in `{artifact_root}/{session_id}/runs.log` and also output to terminal.
+Logs are saved in `{artifact_root}/{session_id}/runs_log.txt` and also output to terminal.
 
 ### Q5: Does it support distributed training?
 

{nextrec-0.4.27 → nextrec-0.4.29}/nextrec_cli_preset/NextRec-CLI_zh.md

@@ -820,7 +820,7 @@ nextrec_logs/
     ├── features_config.pkl        # 特征配置
     ├── model_epoch_1.pt        # 模型检查点
     ├── model_epoch_2.pt
-    └── runs.log               # 训练日志
+    └── runs_log.txt               # 训练日志
 ```
 
 ### Q3: 如何恢复训练？
@@ -829,7 +829,7 @@ nextrec_logs/
 
 ### Q4: 如何查看训练日志？
 
-日志保存在 `{artifact_root}/{session_id}/runs.log`，也会输出到终端。
+日志保存在 `{artifact_root}/{session_id}/runs_log.txt`，也会输出到终端。
 
 ### Q5: 支持分布式训练吗？
 

{nextrec-0.4.27 → nextrec-0.4.29}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "nextrec"
-version = "0.4.27"
+version = "0.4.29"
 description = "A comprehensive recommendation library with match, ranking, and multi-task learning models"
 readme = "README.md"
 requires-python = ">=3.10"

nextrec-0.4.27/nextrec/__version__.py

@@ -1 +0,0 @@
-__version__ = "0.4.27"