Gus Hahn-Powell myedibleenso

The URL structure to download a Google sheet to a csv is https://spreadsheets.google.com/feeds/download/spreadsheets/Export?key=yourDocId&exportFormat=csv&sheet=1

The URL structure to view a Google shet is https://docs.google.com/spreadsheets/d/yourDocId

import pandas as pd
import sqlite3

# pip install "pandas[excel]"

If we want to produce a disjunction of graph traversal patterns (ex. a >rel b OR b >rel c), we currently cannot simply use (a >rel b) | (b >rel c) (this will not compile).

import ai.lum.odinson.ExtractorEngine
import ai.lum.odinson.lucene.search.OdinOrQuery
import ai.lum.odinson.{ Document, Sentence, GraphField, TokensField }

val d1 = Document(
  id = "doc-1",
  metadata = Seq(),

Overview

This short README details the process I followed to perform automatic speech recognition on a 48+ minute audio interview.

1. Convert `m4a` to `mp3`

I Converted m4a to mp3 using ffmpeg. Assuming you have an audio file named interview.m4a to process, ...

ffmpeg -i interview.m4a -codec:a libmp3lame -qscale:a 1 interview.mp3

Using the experimental Odinson REST API

Grab the most recent image

First, pull the most recently pushed image with the :experimental tag:

docker pull lumai/odinson-rest-api:experimental

	# pip install pyfoma>=1
	from pyfoma import FST

	# The Python Finite-State Toolkit (pyfoma)
	# has support for finite-state transducers.
	# FSTs can be used to rewrite symbols in a context dependent manner.
	# These rules can even be weighted.
	# One application of an FST is to convert a written form of a word
	# into IPA (and reverse the process).
	# For languages with a phonologically transparent orthography, this rewrite can be straightforward.

	# tested with ...
	# numpy==1.26.0
	# scikit-learn==1.2.2
	from sklearn import preprocessing
	from sklearn.pipeline import FeatureUnion, Pipeline
	from sklearn.preprocessing import PolynomialFeatures
	from sklearn.feature_extraction.text import CountVectorizer
	from sklearn.linear_model import LogisticRegression
	import numpy as np

	# conda create -n "odinsynth-redux" python=3.9 ipython
	# clone https://github.com/lum-ai/odinson-gateway
	# pip install ".[all]"

	# from the iPython REPL
	from typing import Dict, Any, List, Text
	from odinson.gateway import OdinsonGateway, Document
	import json

	# FIXME: our specification of Odinson documents.

	from typing import (
	Callable,
	Iterable,
	Sequence,
	Tuple,
	TypeAlias
	)
	from collections import Counter
	from itertools import combinations
	from dataclasses import dataclass

	from clu.bridge import odinson
	from clu.bridge import processors
	from clu.bridge import spacy as sp

	from typing import Sequence
	from joblib import Parallel, delayed
	from typing import Dict, Text
	import pydantic
	import spacy
	import pandas as pd

	from __future__ import annotations
	from collections import Counter
	from typing import Dict, List, Optional, Text
	from dataclasses import dataclass

	@dataclass
	class Span:
	start: int
	end: int
	tokens: Optional[List[Text]]

Gus Hahn-Powell myedibleenso

Overview

1. Convert m4a to mp3

Using the experimental Odinson REST API

Grab the most recent image

Launch the REST API

1. Convert `m4a` to `mp3`