Demo.py Run Fails - Githubissues

Issue: Setup and Demo Issues

Instructions and Setup

I've followed the instructions provided in the Quick Setup guide but I seem to have encountered a few issues along the way. While I managed to successfully clone the API-Miner repository and install the required dependencies, I encountered difficulties when setting up the environment to download the BERT model and boot up the demo.

Following along the instruction on Quick Setup

:white_check_mark: 1) Clone API-Miner into your local system

:white_check_mark: 2) Load your dataset

  wget https://github.com/APIs-guru/openapi-directory/blob/main/APIs/openai.com/1.2.0/openapi.yaml

  ls -lrt data/raw/
  total 408
  -rw-r--r-- 1 gitpod gitpod 414736 Aug  7 21:24 openapi.yaml

:warning: 3) Set up your environment

:white_check_mark: pip install all dependencies

:warning: load WordNet library

  gitpod ~ $ ls -lrt ~/nltk_data/
  total 0
  drwxr-xr-x 2 gitpod gitpod 25 Aug 10 07:42 corpora

:eyes: Observation: By default, data is downloaded into the ~/nltk_data directory. However, the instructions lack specificity on the download location.

:white_check_mark: Download BERT-base uncased model

Source: https://storage.googleapis.com/bert_models/2018_10_18/uncased_L-12_H-768_A-12.zip

  tree lib/bert/
  lib/bert/
  ├── bert_config.json
  ├── bert_model.ckpt.data-00000-of-00001
  ├── bert_model.ckpt.index
  ├── bert_model.ckpt.meta
  └── vocab.txt

  0 directories, 5 files

:eyes: Observation: Successfully matched the directory structure as depicted in the provided image

:warning: Download sentence bert model (MiniLM-L6-H384-uncased)

Source A: https://huggingface.co/sentence-transformers/all-MiniLM-L6-v2/tree/main Source B: https://huggingface.co/nreimers/MiniLM-L6-H384-uncased/tree/main

  tree -r lib/sent_bert/
  lib/sent_bert/
  ├── vocab.txt
  ├── train_script.py
  ├── tokenizer.json
  ├── tokenizer_config.json
  ├── tf_model.h5
  ├── special_tokens_map.json
  ├── sentence_bert_config.json
  ├── rust_model.ot
  ├── README.md
  ├── pytorch_model.bin
  ├── modules.json
  ├── data_config.json
  ├── config_sentence_transformers.json
  ├── config.json
  └── 1_Pooling
      └── config.json

  1 directory, 15 files

:eyes: Observation: Neither of the directory structure align with the provided image

:warning: Boot up demo

python demo.py 
Traceback (most recent call last):
  File "/workspace/api-miner/demo.py", line 7, in <module>
    from experiments.utils import transform_endpoint_name_to_features, transform_specs_to_features, retrieve_endpoints_from_database, normalize_scores
ModuleNotFoundError: No module named 'experiments'

:eyes: Observation: The code structure is inconsistent. The mentioned functions are not under the experiments package as expected.

:hammer: Solution Attempt- Revised the script to locate the functions.


# from experiments.utils import transform_endpoint_name_to_features, transform_specs_to_features, retrieve_endpoints_from_database, normalize_scores
# from experiments.setup import initialize_demo

from api_miner.data_processing.utils import testing_func
from api_miner.data_processing.utils import transform_endpoint_name_to_features, transform_specs_to_features, retrieve_endpoints_from_database, normalize_scores
from tests.setup import initialize_demo
from experiments.sample_generator import MaskedSpecGenerator, MangledSpecGenerator, EndpointNameGenerator, UserStudySpecGenerator

However, the problem persists as the mentioned setup and sample_generator scripts were not found in the repository.

jpmorganchase / api-miner

Demo.py Run Fails #3

Issue: Setup and Demo Issues

Instructions and Setup