Spaces:
Paused
Paused
Commit
·
036a85e
1
Parent(s):
ebacf16
update final code
Browse files- text2sql.py +24 -28
text2sql.py
CHANGED
@@ -106,68 +106,64 @@ class ChatBot():
|
|
106 |
def __init__(self) -> None:
|
107 |
os.environ["CUDA_VISIBLE_DEVICES"] = "0"
|
108 |
model_name = "seeklhy/codes-1b"
|
109 |
-
|
110 |
-
# Load tokenizer and model
|
111 |
self.tokenizer = AutoTokenizer.from_pretrained(model_name)
|
112 |
-
self.model = AutoModelForCausalLM.from_pretrained(model_name, device_map="auto", torch_dtype=torch.float16)
|
113 |
-
|
114 |
-
# Set the device for the model (this ensures it's on either GPU or CPU)
|
115 |
-
self.device = self.model.device # This will get the device the model is loaded on (either CUDA or CPU)
|
116 |
-
|
117 |
-
# Define other parameters
|
118 |
self.max_length = 4096
|
119 |
self.max_new_tokens = 256
|
120 |
self.max_prefix_length = self.max_length - self.max_new_tokens
|
121 |
|
122 |
-
#
|
123 |
self.sic = SchemaItemClassifierInference("Roxanne-WANG/LangSQL")
|
124 |
-
|
125 |
-
# Initialize searcher for DB content (Whoosh index)
|
126 |
self.db_id2content_searcher = dict()
|
127 |
for db_id in os.listdir("db_contents_index"):
|
128 |
index_dir = os.path.join("db_contents_index", db_id)
|
|
|
|
|
129 |
if index.exists_in(index_dir):
|
130 |
ix = index.open_dir(index_dir)
|
131 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
132 |
|
133 |
def get_response(self, question, db_id):
|
134 |
-
# Prepare the data for schema filtering
|
135 |
data = {
|
136 |
"text": question,
|
137 |
"schema": copy.deepcopy(self.db_id2schema[db_id]),
|
138 |
"matched_contents": get_matched_contents(question, self.db_id2content_searcher[db_id])
|
139 |
}
|
140 |
-
|
141 |
-
# Filter schema based on predictions
|
142 |
data = filter_schema(data, self.sic, 6, 10)
|
143 |
data["schema_sequence"] = get_db_schema_sequence(data["schema"])
|
144 |
data["content_sequence"] = get_matched_content_sequence(data["matched_contents"])
|
145 |
-
|
146 |
-
# Prepare input sequence for the model
|
147 |
prefix_seq = data["schema_sequence"] + "\n" + data["content_sequence"] + "\n" + data["text"] + "\n"
|
|
|
148 |
|
149 |
-
input_ids = [self.tokenizer.bos_token_id] + self.tokenizer(prefix_seq, truncation=False)["input_ids"]
|
150 |
if len(input_ids) > self.max_prefix_length:
|
|
|
151 |
input_ids = [self.tokenizer.bos_token_id] + input_ids[-(self.max_prefix_length-1):]
|
152 |
attention_mask = [1] * len(input_ids)
|
153 |
|
154 |
-
# Move input tensors to the same device as the model
|
155 |
inputs = {
|
156 |
-
"input_ids": torch.tensor([input_ids], dtype=torch.int64).to(self.device),
|
157 |
-
"attention_mask": torch.tensor([attention_mask], dtype=torch.int64).to(self.device)
|
158 |
}
|
159 |
-
|
160 |
-
|
161 |
with torch.no_grad():
|
162 |
generate_ids = self.model.generate(
|
163 |
**inputs,
|
164 |
-
max_new_tokens=self.max_new_tokens,
|
165 |
-
num_beams=4,
|
166 |
-
num_return_sequences=4
|
167 |
)
|
168 |
|
169 |
-
|
170 |
-
generated_sqls = self.tokenizer.batch_decode(generate_ids[:, len(input_ids):], skip_special_tokens=True, clean_up_tokenization_spaces=False)
|
171 |
final_generated_sql = None
|
172 |
for generated_sql in generated_sqls:
|
173 |
execution_error = check_sql_executability(generated_sql, os.path.join("databases", db_id, db_id + ".sqlite"))
|
|
|
106 |
def __init__(self) -> None:
|
107 |
os.environ["CUDA_VISIBLE_DEVICES"] = "0"
|
108 |
model_name = "seeklhy/codes-1b"
|
|
|
|
|
109 |
self.tokenizer = AutoTokenizer.from_pretrained(model_name)
|
110 |
+
self.model = AutoModelForCausalLM.from_pretrained(model_name, device_map = "auto", torch_dtype = torch.float16)
|
|
|
|
|
|
|
|
|
|
|
111 |
self.max_length = 4096
|
112 |
self.max_new_tokens = 256
|
113 |
self.max_prefix_length = self.max_length - self.max_new_tokens
|
114 |
|
115 |
+
# Directly loading the model from Hugging Face
|
116 |
self.sic = SchemaItemClassifierInference("Roxanne-WANG/LangSQL")
|
|
|
|
|
117 |
self.db_id2content_searcher = dict()
|
118 |
for db_id in os.listdir("db_contents_index"):
|
119 |
index_dir = os.path.join("db_contents_index", db_id)
|
120 |
+
|
121 |
+
# Open existing Whoosh index directory
|
122 |
if index.exists_in(index_dir):
|
123 |
ix = index.open_dir(index_dir)
|
124 |
+
# keep a searcher around for querying
|
125 |
+
self.db_id2content_searcher[db_id] = ix.searcher()
|
126 |
+
else:
|
127 |
+
raise ValueError(f"No Whoosh index found for '{db_id}' at '{index_dir}'")
|
128 |
+
|
129 |
+
self.db_ids = sorted(os.listdir("databases"))
|
130 |
+
self.db_id2schema = get_db_id2schema("databases", "data/tables.json")
|
131 |
+
self.db_id2ddl = get_db_id2ddl("databases")
|
132 |
|
133 |
def get_response(self, question, db_id):
|
|
|
134 |
data = {
|
135 |
"text": question,
|
136 |
"schema": copy.deepcopy(self.db_id2schema[db_id]),
|
137 |
"matched_contents": get_matched_contents(question, self.db_id2content_searcher[db_id])
|
138 |
}
|
|
|
|
|
139 |
data = filter_schema(data, self.sic, 6, 10)
|
140 |
data["schema_sequence"] = get_db_schema_sequence(data["schema"])
|
141 |
data["content_sequence"] = get_matched_content_sequence(data["matched_contents"])
|
142 |
+
|
|
|
143 |
prefix_seq = data["schema_sequence"] + "\n" + data["content_sequence"] + "\n" + data["text"] + "\n"
|
144 |
+
print(prefix_seq)
|
145 |
|
146 |
+
input_ids = [self.tokenizer.bos_token_id] + self.tokenizer(prefix_seq , truncation = False)["input_ids"]
|
147 |
if len(input_ids) > self.max_prefix_length:
|
148 |
+
print("the current input sequence exceeds the max_tokens, we will truncate it.")
|
149 |
input_ids = [self.tokenizer.bos_token_id] + input_ids[-(self.max_prefix_length-1):]
|
150 |
attention_mask = [1] * len(input_ids)
|
151 |
|
|
|
152 |
inputs = {
|
153 |
+
"input_ids": torch.tensor([input_ids], dtype = torch.int64).to(self.model.device),
|
154 |
+
"attention_mask": torch.tensor([attention_mask], dtype = torch.int64).to(self.model.device)
|
155 |
}
|
156 |
+
input_length = inputs["input_ids"].shape[1]
|
157 |
+
|
158 |
with torch.no_grad():
|
159 |
generate_ids = self.model.generate(
|
160 |
**inputs,
|
161 |
+
max_new_tokens = self.max_new_tokens,
|
162 |
+
num_beams = 4,
|
163 |
+
num_return_sequences = 4
|
164 |
)
|
165 |
|
166 |
+
generated_sqls = self.tokenizer.batch_decode(generate_ids[:, input_length:], skip_special_tokens = True, clean_up_tokenization_spaces = False)
|
|
|
167 |
final_generated_sql = None
|
168 |
for generated_sql in generated_sqls:
|
169 |
execution_error = check_sql_executability(generated_sql, os.path.join("databases", db_id, db_id + ".sqlite"))
|