pip install paperextpdftotext from https://poppler.freedesktop.org/ is required to run this utility.
usage: download-convert [-h] [--cache-dir DIR] JSON
Utility to download and convert a list of papers' pdfs -> "txts.
stdout will contain the list of files successfully converted separated by '\n'.
positional arguments:
JSON Paperoni json output of papers to download and convert pdfs -> txts
options:
-h, --help show this help message and exit
--cache-dir DIR Directory to store downloaded and converted pdfs -> txts
Example:
$ download-convert paperoni-2024-07-04.json
[DEBUG]
data/cache/arxiv/1901.07186.txt
data/cache/arxiv/1906.05433.txt
...
data/cache/html/874f823e6462acbbb07cc57d32e09217.txt
data/cache/html/8a46fcbc0c34ea85102920cba7039290.txt
...
data/cache/openreview/0k_DN90uWF.txt
data/cache/openreview/2Q8TZWAHv4.txt
...
data/cache/pdf/80c62591b54231aa42e4418fe3d45e8f.txt
data/cache/pdf/81709b4783324a59fd2632ee694e9071.txt
...
Successfully downloaded and converted 587 out of 867 papers
arxiv:455/455
html:3/3
openreview:52/52
pdf:77/145
$ download-convert paperoni-2024-07-04.json > data/query_set.txt
[DEBUG]
Successfully downloaded and converted 587 out of 867 papers
arxiv:455/455
html:3/3
openreview:52/52
pdf:77/145An OpenAI API Key is required to run this utility:
export OPENAI_API_KEY=sk-proj-XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXusage: query [-h] [--papers [PAPERS ...]] [--input TXT]
Utility to query Chat-GPT on papers
options:
-h, --help show this help message and exit
--papers [PAPERS ...]
Papers to analyse
--input TXT List of papers to analyse
Example:
$ query --input data/query_set.txt > query.out 2> query.errusage: parse-validation-errors [-h] TXT
Utility to parse and gather stats on query logs
positional arguments:
TXT query stdout file to parse
options:
-h, --help show this help message and exit
Example:
$ parse-validation-errors query.err
Per paper errors
================
Failures count for paper: 1
---------------------------
primary_research_field.aliases 1 ~1
primary_research_field.name 1 ~1
sub_research_fields.*.aliases 2 ~1
sub_research_fields.*.name 2 ~1
[...]
Failures count for paper: 2
---------------------------
datasets 1 ~1
models.* 13 ~1
sub_research_fields.*.aliases 1 ~1
sub_research_fields.*.name 1 ~1
libraries.*.aliases 1 ~1
libraries.*.referenced_paper_title 1 ~1
libraries.*.role 1 ~1
[...]
Failures count for paper: 2
---------------------------
Generic Error:
Invalid JSON: control character (\u0000-\u001F) found while parsing a string at line 1 column 2885 [...]
For further information visit https://errors.pydantic.dev/2.7/v/json_invalid,)
Generic Error:
Invalid JSON: control character (\u0000-\u001F) found while parsing a string at line 1 column 2885 [...]
For further information visit https://errors.pydantic.dev/2.7/v/json_invalid,)
[...]
Error type involved in a paper validation
=========================================
Generic Error 2 /62
datasets 5 /62
libraries.*.aliases 10 /62
libraries.*.referenced_paper_title 9 /62
libraries.*.role 9 /62
models.* 3 /62
primary_research_field.aliases 40 /62
primary_research_field.name 42 /62
sub_research_fields.*.aliases 32 /62
sub_research_fields.*.name 37 /62
Queries Stats
=============
Requests cnt 204
Errors cnt 62
1 failure(s) 46
2 failure(s) 8usage: analysis [-h] [--papers [PAPERS ...]] [--input TXT]
Utility to analyses Chat-GPT responses on papers
Confidence and multi-label confidence matrices will be dumped into data/analysis
options:
-h, --help show this help message and exit
--papers [PAPERS ...]
Papers to analyse
--input TXT List of papers to analyse
Example:
$ analysispaperext is distributed under the terms of the MIT license.