Tutorial 24_Building_Chat_App.ipynb is not working

[dkbs12]

Hello,
I need your help on 'Tutorial: Building a Conversational Chat App'.
When I tried '24_Building_Chat_App.ipynb' on my Colab, I faced such error messages as below.

The first one was about 'Enabling Telemetry' and I tried just as same as written on the Colab.
But I found the error message 'TypeError: catching classes that do not inherit from BaseException is not allowed' responsed to below codings;

from haystack.telemetry import tutorial_running
tutorial_running(24)

Second, when I tried to create PromptNode using below codes;

from haystack.nodes import PromptNode
model_name = "OpenAssistant/oasst-sft-1-pythia-12b"
prompt_node = PromptNode(model_name, api_key=model_api_key, max_length=256)

I also faced 'MissingDependencyException' and 'TypeError: catching classes that do not inherit from BaseException is not allowed'.
The details are as below;

╭─────────────────────────────── Traceback (most recent call last) ────────────────────────────────╮
│ /usr/local/lib/python3.10/dist-packages/haystack/init.py:84 in is_imported │
│ │
│ 81 │ """ │
│ 82 │ module = sys.modules.get(module_name) │
│ 83 │ try: │
│ ❱ 84 │ │ return bool(module) and not isinstance(module, FakeModule) │
│ 85 │ except MissingOptionalDependency: │
│ 86 │ │ # isinstance() raises MissingOptionalDependency: fake module │
│ 87 │ │ pass │
│ │
│ /usr/local/lib/python3.10/dist-packages/generalimport/fake_module.py:69 in error_func │
│ │
│ 66 │ │
│ 67 │ def error_func(self, _caller: str, *args, **kwargs): │
│ 68 │ │ error_pars = ErrorPars(name=self.name, trigger=self.trigger, caller=_caller, cat │
│ ❱ 69 │ │ return self._error_func(error_pars=error_pars) │
│ 70 │ │
│ 71 │ @classmethod │
│ 72 │ def error_func_class(cls, _caller: str, *args, **kwargs): │
│ │
│ /usr/local/lib/python3.10/dist-packages/generalimport/fake_module.py:65 in _error_func │
│ │
│ 62 │ │ msg = cls._get_error_message(error_pars=error_pars) │
│ 63 │ │ │
│ 64 │ │ logger.debug(msg=msg) │
│ ❱ 65 │ │ raise MissingDependencyException(msg=msg) │
│ 66 │ │
│ 67 │ def error_func(self, _caller: str, *args, **kwargs): │
│ 68 │ │ error_pars = ErrorPars(name=self.name, trigger=self.trigger, caller=_caller, cat │
╰──────────────────────────────────────────────────────────────────────────────────────────────────╯
MissingDependencyException: Optional dependency fitz was used but it isn't installed. Triggered by 'bool'.

During handling of the above exception, another exception occurred:

╭─────────────────────────────── Traceback (most recent call last) ────────────────────────────────╮
│ in <cell line: 1>:1 │
│ │
│ /usr/local/lib/python3.10/dist-packages/haystack/init.py:93 in │
│ │
│ 90 │
│ 91 from haystack.schema import Document, Answer, Label, MultiLabel, Span, EvaluationResult, │
│ 92 from haystack.nodes.base import BaseComponent │
│ ❱ 93 from haystack.pipelines.base import Pipeline │
│ 94 from haystack.environment import set_pytorch_secure_model_loading │
│ 95 from haystack.mmh3 import hash128 │
│ 96 │
│ │
│ /usr/local/lib/python3.10/dist-packages/haystack/pipelines/init.py:1 in │
│ │
│ ❱ 1 from haystack.pipelines.base import Pipeline, RootNode │
│ 2 from haystack.pipelines.ray import RayPipeline │
│ 3 from haystack.pipelines.standard_pipelines import ( │
│ 4 │ BaseStandardPipeline, │
│ │
│ /usr/local/lib/python3.10/dist-packages/haystack/pipelines/base.py:33 in │
│ │
│ 30 from networkx.drawing.nx_agraph import to_agraph │
│ 31 │
│ 32 from haystack import version │
│ ❱ 33 from haystack.modeling.evaluation.metrics import semantic_answer_similarity │
│ 34 from haystack.modeling.evaluation.squad import compute_f1 as calculate_f1_str │
│ 35 from haystack.modeling.evaluation.squad import compute_exact as calculate_em_str │
│ 36 from haystack.pipelines.config import ( │
│ │
│ /usr/local/lib/python3.10/dist-packages/haystack/modeling/evaluation/init.py:1 in │
│ │
│ ❱ 1 from haystack.modeling.evaluation.eval import Evaluator │
│ 2 │
│ │
│ /usr/local/lib/python3.10/dist-packages/haystack/modeling/evaluation/eval.py:10 in │
│ │
│ 7 import numpy as np │
│ 8 from tqdm.auto import tqdm │
│ 9 │
│ ❱ 10 from haystack.modeling.evaluation.metrics import compute_metrics, compute_report_metrics │
│ 11 from haystack.modeling.model.adaptive_model import AdaptiveModel │
│ 12 from haystack.modeling.model.biadaptive_model import BiAdaptiveModel │
│ 13 from haystack.modeling.model.optimization import WrappedDataParallel │
│ │
│ /usr/local/lib/python3.10/dist-packages/haystack/modeling/evaluation/metrics.py:9 in │
│ │
│ 6 from sentence_transformers import CrossEncoder, SentenceTransformer │
│ 7 from transformers import AutoConfig │
│ 8 │
│ ❱ 9 from haystack.modeling.model.prediction_head import PredictionHead │
│ 10 from haystack.modeling.utils import flatten_list │
│ 11 │
│ 12 │
│ │
│ /usr/local/lib/python3.10/dist-packages/haystack/modeling/model/prediction_head.py:17 in │
│ │
│ │
│ 14 from haystack.modeling.data_handler.samples import SampleBasket │
│ 15 from haystack.modeling.model.predictions import QACandidate, QAPred │
│ 16 from haystack.modeling.utils import try_get, all_gather_list │
│ ❱ 17 from haystack.utils.scipy_utils import expit │
│ 18 │
│ 19 │
│ 20 logger = logging.getLogger(name) │
│ │
│ /usr/local/lib/python3.10/dist-packages/haystack/utils/init.py:2 in │
│ │
│ 1 from haystack.utils.reflection import args_to_kwargs │
│ ❱ 2 from haystack.utils.preprocessing import convert_files_to_docs, tika_convert_files_to_do │
│ 3 from haystack.utils.import_utils import fetch_archive_from_http │
│ 4 from haystack.utils.cleaning import clean_wiki_text │
│ 5 from haystack.utils.doc_store import launch_es, launch_opensearch, launch_weaviate, stop │
│ │
│ /usr/local/lib/python3.10/dist-packages/haystack/utils/preprocessing.py:7 in │
│ │
│ 4 import logging │
│ 5 from pathlib import Path │
│ 6 │
│ ❱ 7 from haystack.nodes.file_converter import BaseConverter, DocxToTextConverter, PDFToTextC │
│ 8 from haystack.schema import Document │
│ 9 │
│ 10 │
│ │
│ /usr/local/lib/python3.10/dist-packages/haystack/nodes/init.py:9 in │
│ │
│ 6 from haystack.nodes.document_classifier import BaseDocumentClassifier, TransformersDocum │
│ 7 from haystack.nodes.extractor import EntityExtractor, simplify_ner_for_qa │
│ 8 from haystack.nodes.file_classifier import FileTypeClassifier │
│ ❱ 9 from haystack.nodes.file_converter import ( │
│ 10 │ BaseConverter, │
│ 11 │ DocxToTextConverter, │
│ 12 │ ImageToTextConverter, │
│ │
│ /usr/local/lib/python3.10/dist-packages/haystack/nodes/file_converter/init.py:17 in │
│ │
│ 14 # Try to use PyMuPDF, if not available fall back to xpdf │
│ 15 from haystack.nodes.file_converter.pdf import PDFToTextConverter │
│ 16 │
│ ❱ 17 if not is_imported("fitz"): │
│ 18 │ from haystack.nodes.file_converter.pdf_xpdf import PDFToTextConverter # type: ignor │
│ 19 │
│ 20 │
│ │
│ /usr/local/lib/python3.10/dist-packages/haystack/init.py:85 in is_imported │
│ │
│ 82 │ module = sys.modules.get(module_name) │
│ 83 │ try: │
│ 84 │ │ return bool(module) and not isinstance(module, FakeModule) │
│ ❱ 85 │ except MissingOptionalDependency: │
│ 86 │ │ # isinstance() raises MissingOptionalDependency: fake module │
│ 87 │ │ pass │
│ 88 │ return False │
╰──────────────────────────────────────────────────────────────────────────────────────────────────╯
TypeError: catching classes that do not inherit from BaseException is not allowed

I guess all these problems might come from dependency conflicts.
When I install 'upgrade pip' and 'farm-haystack[colab]', I found the error message as below;

"ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.
tensorflow 2.12.0 requires protobuf!=4.21.0,!=4.21.1,!=4.21.2,!=4.21.3,!=4.21.4,!=4.21.5,<5.0.0dev,>=3.20.3, but you have protobuf 3.20.2 which is incompatible.
tensorflow-metadata 1.13.1 requires protobuf<5,>=3.20.3, but you have protobuf 3.20.2 which is incompatible."

Now, I don't have any idea to solve these problems and need a help from the experts.
Thanks.

[agario8864]

Having same issue here, manage to solve it using different Tensor Flow version. Try this:

!pip install TensorFlow==2.12

[dkbs12]

Hello,
Thank you for your help.
I've tried as you advised above, but it doesn't work.
I found the same error messages as I reported two days ago.

When I tried different Tensor Flow version as below;

!pip install TensorFlow==2.12

the output was responded as below;

ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.
farm-haystack 1.17.0 requires protobuf<=3.20.2, but you have protobuf 4.23.2 which is incompatible.
Successfully installed protobuf-4.23.2

[agario8864]

what about:

!pip install protobuf==3.20.2

It seems like the protobuf package you are using is not compatible with haystack.

[TuanaCelik]

Hey @dkbs12 and @agario8864 thanks for reporting this. We're looking into it and we'll update the discussion 🙏

[TuanaCelik]

@ZanSara is pushing a fix for this. Seems the same issue occurs on all tutorials because of a breaking change in generalimports library: #5074
Hopefully there should be a patch release soon.

[ZanSara]

Hey @dkbs12 and @agario8864, downgrading generalimport should fix the issue. Try running pip install generalimport==0.3.1 and let us know if that fixes your issue. Sorry for the trouble!

[agario8864]

@ZanSara Yes, it works now, tq.

[ZanSara]

Thank you for the confirmation 😊

[dkbs12]

Dear ZanSara, Thank you for your quick reply and action. Now the problem is solved and it works. Could you reflect this solution on the existing tutorial for the people who don't know this matter? Thanks.

…

-----Original Message----- From: ***@***.***> To: ***@***.***>; Cc: ***@***.***>; ***@***.***>; Sent: 2023-06-05 (월) 17:23:18 (GMT+09:00) Subject: Re: [deepset-ai/haystack] Tutorial 24_Building_Chat_App.ipynb is not working (Discussion #5072) Hey @dkbs12 and @agario8864, downgrading generalimport should fix the issue. Try running pip install generalimport==0.3.1 and let us know if that fixes your issue. Sorry for the trouble! — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: ***@***.***>

[ZanSara]

We released a patch to address the issue: 1.17.1. The tutorial works now.