Fix xgb string data contains slash (#2869)

* fix xgb string data contains slash

* fix ci

Author: typhoonzero

python/runtime/local/xgboost_submitter/predict.py

@@ -23,7 +23,7 @@
 from runtime.feature.derivation import get_ordered_field_descs
 from runtime.feature.field_desc import DataType
 from runtime.model.model import Model
-from runtime.xgboost.dataset import xgb_dataset
+from runtime.xgboost.dataset import DMATRIX_FILE_SEP, xgb_dataset
 
 
 def pred(datasource, select, result_table, pred_label_name, model):
@@ -148,8 +148,8 @@ def _store_predict_result(preds, result_table, result_column_names,
                     break
 
                 row = [
-                    item for i, item in enumerate(line.strip().split("/"))
-                    if i != train_label_idx
+                    item for i, item in enumerate(line.strip().split(
+                        DMATRIX_FILE_SEP)) if i != train_label_idx
                 ]
                 row.append(str(preds[line_no]))
                 w.write(row)

python/runtime/xgboost/dataset.py

@@ -25,6 +25,8 @@
 from scipy.sparse import vstack
 from sklearn.datasets import load_svmlight_file, load_svmlight_files
 
+DMATRIX_FILE_SEP = "\t"
+
 
 def xgb_dataset(datasource,
                 fn,
@@ -130,7 +132,8 @@ def dump_dmatrix(filename,
                                                  feature_metas)
 
             if raw_data_fid is not None:
-                raw_data_fid.write("/".join([str(r) for r in row]) + "\n")
+                raw_data_fid.write(
+                    DMATRIX_FILE_SEP.join([str(r) for r in row]) + "\n")
 
             if transform_fn:
                 features = transform_fn(features)
@@ -163,7 +166,7 @@ def dump_dmatrix(filename,
             if has_label:
                 row_data = [str(label)] + row_data
 
-            f.write("\t".join(row_data) + "\n")
+            f.write(DMATRIX_FILE_SEP.join(row_data) + "\n")
             row_id += 1
             # batch_size == None means use all data in generator
             if batch_size is None:

python/runtime/xgboost/evaluate.py

@@ -17,7 +17,7 @@
 from runtime import db
 from runtime.dbapi.paiio import PaiIOConnection
 from runtime.model.metadata import load_metadata
-from runtime.xgboost.dataset import xgb_dataset
+from runtime.xgboost.dataset import DMATRIX_FILE_SEP, xgb_dataset
 
 SKLEARN_METRICS = [
     'accuracy_score',
@@ -118,7 +118,7 @@ def evaluate_and_store_result(bst, dpred, feature_file_id, validation_metrics,
 
     y_test_list = []
     for line in feature_file_read:
-        row = [i for i in line.strip().split("\t")]
+        row = [i for i in line.strip().split(DMATRIX_FILE_SEP)]
         # DMatrix store label in the first column
         if label_meta["dtype"] == "float32":
             label = float(row[0])

python/runtime/xgboost/predict.py

@@ -16,7 +16,7 @@
 from runtime import db
 from runtime.dbapi.paiio import PaiIOConnection
 from runtime.model.metadata import load_metadata
-from runtime.xgboost.dataset import xgb_dataset
+from runtime.xgboost.dataset import DMATRIX_FILE_SEP, xgb_dataset
 
 DEFAULT_PREDICT_BATCH_SIZE = 10000
 
@@ -123,7 +123,8 @@ def predict_and_store_result(bst, dpred, feature_file_id, model_params,
             # FIXME(typhoonzero): how to output columns that are not used
             # as features, like ids?
             row = [
-                item for i, item in enumerate(line.strip().split("/"))
+                item
+                for i, item in enumerate(line.strip().split(DMATRIX_FILE_SEP))
                 if i != train_label_index
             ]
             row.append(str(preds[line_no]))