search

neulab / prompt2model

prompt2model - Generate Deployable Models from Natural Language Instructions
Apache License 2.0
1.93k stars 166 forks source link

TypeError: 'type' object is not subscriptable #325

Open Souryadipstan opened 10 months ago

Souryadipstan commented 10 months ago

Getting this while importing OpenAIInstructionParser, TaskType

TypeError Traceback (most recent call last) Cell In[1], line 1 ----> 1 from prompt2model.prompt_parser import OpenAIInstructionParser, TaskType 3 prompt_spec = OpenAIInstructionParser(task_type=TaskType.TEXT_GENERATION) 4 prompt_spec.parse_from_prompt(prompt)

File ~/.pyenv/versions/3.8.17/lib/python3.8/site-packages/prompt2model/prompt_parser/init.py:3 1 """Import PromptSpec classes.""" 2 from prompt2model.prompt_parser.base import PromptSpec, TaskType ----> 3 from prompt2model.prompt_parser.instr_parser import OpenAIInstructionParser 4 from prompt2model.prompt_parser.mock import MockPromptSpec 6 all = ( 7 "PromptSpec", 8 "TaskType", 9 "MockPromptSpec", 10 "OpenAIInstructionParser", 11 )

File ~/.pyenv/versions/3.8.17/lib/python3.8/site-packages/prompt2model/prompt_parser/instr_parser.py:15 10 from prompt2model.prompt_parser.base import PromptSpec, TaskType 12 from prompt2model.prompt_parser.instr_parser_prompt import ( # isort: split 13 construct_prompt_for_instruction_parsing, 14 ) ---> 15 from prompt2model.utils import ( 16 OPENAI_ERRORS, 17 ChatGPTAgent, 18 get_formatted_logger, 19 handle_openai_error, 20 ) 22 logger = get_formatted_logger("PromptParser") 24 os.environ["TOKENIZERS_PARALLELISM"] = "false"

File ~/.pyenv/versions/3.8.17/lib/python3.8/site-packages/prompt2model/utils/init.py:10 3 from prompt2model.utils.openai_tools import ( 4 OPENAI_ERRORS, 5 ChatGPTAgent, 6 count_tokens_from_string, 7 handle_openai_error, 8 ) 9 from prompt2model.utils.rng import seed_generator ---> 10 from prompt2model.utils.tevatron_utils import encode_text, retrieve_objects 12 all = ( # noqa: F401 13 "ChatGPTAgent", 14 "encode_text", (...) 20 "get_formatted_logger", 21 )

File ~/.pyenv/versions/3.8.17/lib/python3.8/site-packages/prompt2model/utils/tevatron_utils/init.py:3 1 """Import Tevatron utility functions.""" 2 from prompt2model.utils.tevatron_utils.encode import encode_text ----> 3 from prompt2model.utils.tevatron_utils.retrieve import retrieve_objects 5 all = ("encode_text", "retrieve_objects")

File ~/.pyenv/versions/3.8.17/lib/python3.8/site-packages/prompt2model/utils/tevatron_utils/retrieve.py:11 4 import numpy as np 5 from tevatron.faiss_retriever import BaseFaissIPRetriever 8 def retrieve_objects( 9 query_vector: np.ndarray, 10 encoded_datasets_path: str, ---> 11 document_names: list[str], 12 depth: int, 13 ) -> list[tuple[str, float]]: 14 """Return a ranked list of object indices and their scores. 15 16 Args: (...) 22 Ranked list of object names and their inner product similarity to the query. 23 """ 24 assert query_vector.shape[0] == 1, "Only a single query vector is expected."

TypeError: 'type' object is not subscriptable

neubig commented 10 months ago

Thank you for finding this and sending a fix @Souryadipstan !

This should have been caught through our automatic tests, but it wasn't for some reason. So I'd like to keep it open even after the original PR is merged so we can add tests.