def test_nlp_file_writer():
"""test nlp file writer using shard api"""
schema_json = {
"id": {
"type": "string"
},
"label": {
"type": "number"
},
"rating": {
"type": "number"
},
"input_ids": {
"type": "array",
"items": {
"type": "number"
}
},
"input_mask": {
"type": "array",
"items": {
"type": "number"
}
},
"segment_ids": {
"type": "array",
"items": {
"type": "number"
}
}
}
data = list(
get_nlp_data("../data/mindrecord/testAclImdbData/pos",
"../data/mindrecord/testAclImdbData/vocab.txt", 10))
header = ShardHeader()
schema = header.build_schema(schema_json, ["segment_ids"], "nlp_schema")
schema_id = header.add_schema(schema)
assert schema_id == 0, 'failed on adding schema'
index_fields_list = ["id", "rating"]
ret = header.add_index_fields(index_fields_list)
assert ret == SUCCESS, 'failed on adding index fields.'
writer = ShardWriter()
paths = ["{}{}".format(NLP_FILE_NAME, x) for x in range(FILES_NUM)]
ret = writer.open(paths)
assert ret == SUCCESS, 'failed on opening files.'
writer.set_header_size(1 << 14)
writer.set_page_size(1 << 15)
ret = writer.set_shard_header(header)
assert ret == SUCCESS, 'failed on setting header.'
ret = writer.write_raw_nlp_data({schema_id: data})
assert ret == SUCCESS, 'failed on writing raw data.'
ret = writer.commit()
assert ret == SUCCESS, 'failed on committing.'
generator = ShardIndexGenerator(os.path.realpath(paths[0]))
generator.build()
generator.write_to_db()
def test_mkv_file_writer():
"""test mkv file writer using shard api"""
data = get_mkv_data("../data/mindrecord/testVehPerData/")
schema_json = {
"file_name": {
"type": "string"
},
"id": {
"type": "number"
},
"prelabel": {
"type": "string"
}
}
header = ShardHeader()
img_schema = header.build_schema(schema_json, ["data"], "img_schema")
schema_id = header.add_schema(img_schema)
assert schema_id == 0, 'failed on building schema.'
index_fields_list = ["id", "file_name"]
ret = header.add_index_fields(index_fields_list)
assert ret == SUCCESS, 'failed on adding index fields.'
writer = ShardWriter()
paths = ["{}{}".format(MKV_FILE_NAME, x) for x in range(FILES_NUM)]
ret = writer.open(paths)
assert ret == SUCCESS, 'failed on opening files.'
writer.set_header_size(1 << 24)
writer.set_page_size(1 << 25)
ret = writer.set_shard_header(header)
assert ret == SUCCESS, 'failed on setting header.'
ret = writer.write_raw_cv_data({schema_id: data})
assert ret == SUCCESS, 'failed on writing raw data.'
ret = writer.commit()
assert ret == SUCCESS, 'failed on committing.'
generator = ShardIndexGenerator(os.path.realpath(paths[0]))
generator.build()
generator.write_to_db()
def test_file_writer_fail_add_index():
"""test file writer, read when failed on adding index."""
data_raw = get_data("../data/mindrecord/testImageNetData/")
schema_json = {
"file_name": {
"type": "string"
},
"label": {
"type": "number"
}
}
header = ShardHeader()
schema = header.build_schema(schema_json, ["data"], "img") # create schema
schema_id = header.add_schema(schema) # add schema
with pytest.raises(TypeError, match="missing 1 "):
ret = header.add_index_fields()
assert ret == FAILED
with pytest.raises(MRMAddIndexError):
index_fields = []
ret = header.add_index_fields(index_fields)
assert ret == FAILED
file_name = os.path.join(os.getcwd(),
"test_001.mindrecord") # set output filename
writer = ShardWriter() # test_file_writer
ret = writer.open([file_name])
assert ret == SUCCESS, 'failed on opening files.'
ret = writer.set_shard_header(header) # write header
assert ret == SUCCESS, 'failed on setting header.'
ret = writer.write_raw_cv_data({schema_id: data_raw})
assert ret == SUCCESS, 'failed on writing raw data.'
ret = writer.commit() # commit data
assert ret == SUCCESS, "commit failed"
# ShardIndexGenerator
generator = ShardIndexGenerator(os.path.realpath(file_name))
generator.build()
generator.write_to_db()
reader = ShardReader()
ret = reader.open(file_name)
reader.launch()
index = 0
_, blob_fields = reader.get_blob_fields()
iterator = reader.get_next()
while iterator:
for blob, raw in iterator:
raw[blob_fields[0]] = bytes(blob)
logger.info("#item{}: {}".format(index, raw))
index += 1
iterator = reader.get_next()
reader.finish()
reader.close()
os.remove("{}".format(file_name))
os.remove("{}.db".format(file_name))
def test_mkv_file_writer_with_exactly_schema():
"""test mkv file writer using shard api"""
header = ShardHeader()
img_schema_json = {
"annotation_name": {
"type": "array",
"items": {
"type": "string"
}
},
"annotation_pose": {
"type": "array",
"items": {
"type": "string"
}
},
"annotation_truncated": {
"type": "array",
"items": {
"type": "string"
}
},
"annotation_difficult": {
"type": "array",
"items": {
"type": "string"
}
},
"annotation_xmin": {
"type": "array",
"items": {
"type": "number"
}
},
"annotation_ymin": {
"type": "array",
"items": {
"type": "number"
}
},
"annotation_xmax": {
"type": "array",
"items": {
"type": "number"
}
},
"annotation_ymax": {
"type": "array",
"items": {
"type": "number"
}
},
"metadata_width": {
"type": "number"
},
"metadata_height": {
"type": "number"
},
"metadata_depth": {
"type": "number"
},
"img_path": {
"type": "string"
},
"annotation_path": {
"type": "string"
}
}
img_schema = header.build_schema(img_schema_json, ["data"], "image_schema")
schema_id = header.add_schema(img_schema)
assert schema_id == 0, 'failed on building schema.'
writer = ShardWriter()
paths = ["{}{}".format(MKV_FILE_NAME, x) for x in range(1)]
ret = writer.open(paths)
assert ret == SUCCESS, 'failed on opening files.'
writer.set_header_size(1 << 24)
writer.set_page_size(1 << 25)
image_bytes = bytes("it's a image picutre", encoding="utf8")
data = []
data.append({
"annotation_name": ["xxxxxxxxxx.jpg"],
"annotation_pose": ["hahahahah"],
"annotation_truncated": ["1"],
"annotation_difficult": ["0"],
"annotation_xmin": [100],
"annotation_ymin": [200],
"annotation_xmax": [300],
"annotation_ymax": [400],
"metadata_width": 333,
"metadata_height": 222,
"metadata_depth": 3,
"img_path": "/tmp/",
"annotation_path": "/tmp/annotation",
"data": image_bytes
})
data.append({
"annotation_name": ["xxxxxxxxxx.jpg"],
"annotation_pose": ["hahahahah"],
"annotation_truncated": ["1"],
"annotation_difficult": ["0"],
"annotation_xmin": [100],
"annotation_ymin": [200],
"annotation_xmax": [300],
"annotation_ymax": [400],
"metadata_width": 333,
"metadata_height": 222,
"metadata_depth": 3,
"img_path": "/tmp/",
"annotation_path": "/tmp/annotation",
"data": image_bytes
})
ret = writer.set_shard_header(header)
assert ret == SUCCESS, 'failed on setting header.'
ret = writer.write_raw_cv_data({schema_id: data})
assert ret == SUCCESS, 'failed on writing raw data.'
ret = writer.commit()
assert ret == SUCCESS, 'failed on committing.'
generator = ShardIndexGenerator(os.path.realpath(paths[0]))
generator.build()
generator.write_to_db()