KO | EN

GitLite

Minimalist Open Source Repository Search Engine

gitlite — search

#developer-tools #typescript #claude-code #codex #dsh #dsh-plugin #ai-agent #ai-agents #deepseek-harness #agent-skills #macos #cli

openalexnet

★ 27 Open GitHub ↗

OpenAlex Networks is a helper library to process and obtain data from the OpenAlex dataset via API. It also provides functionality to generate citation and coauthorship networks from queries.

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# OpenAlex Networks (openalexnet)
OpenAlex Networks is a helper library and standalone command-line application to process and obtain data from the [OpenAlex](https://openalex.org) dataset via API. It also provides functionality to generate citation and coauthorship networks from queries.

## [Installation](#installation)

Install using pip

```bash
pip install openalexnet
```

or from source:
```bash
pip git+https://github.com/filipinascimento/openalexnet.git
```

## [Usage as command-line application](#usage-as-command-line-application)
After installing openalexnet, you can use the command:
```bash
python -m openalexnet
```
or simply 
```bash
openalexnet
```
This should print a help message with the available commands and options.

You can make your first query by using:
```bash
openalexnet -t works -f "author.id:A2420755856,is_paratext:false,type:journal-article" -s "complex" -r "cited_by_count:desc" -o works.jsonl -c citation_network.gml -a coauthorship_network.gml
```
This will get all the journal articles from H. Eugene Stanley (A2420755856) with the word "complex" and sorted by the number of citations (in descending order).

For more details about the interface, check the following sections.

### [Querying the OpenAlex API](#querying-the-openalex-api)
The queries have four main parameters:
 - `entitytype` (`-t`): Type of entity to be retrieved from the OpenAlex API. Can be one of the following: `works`, `institutions`, `authors`, `concepts` or `venues`
 - `filter` (`-f`): Comma-separated filter entries formatted as `<key>:<value>` to be used in the OpenAlex API call. Only results passing the filter will be retrieved. See https://docs.openalex.org/how-to-use-the-api/get-lists-of-entities/filter-entity-lists for more information. Defaults to `""` (or no filter). Example: `-f "type:journal-article,author.id:A2420755856"`.
 - `search` (`-s`): Search string to be used in the OpenAlex API call. Only results matching the search string (in the title, abstract, or fulltext) will be retrieved. See https://docs.openalex.org/how-to-use-the-api/get-lists-of-entities/search-entities for more information. Defaults to `""` (or no search). Example: `-s "complex networks"`.
 - `sort` (`-r`): Comma-separated sort entries formatted as `<key>[:desc]` to be used in the OpenAlex API call. See https://docs.openalex.org/how-to-use-the-api/get-lists-of-entities/sort-entity-lists for more information. Defaults to `""` (or no sort). Example: `-r "cited_by_count:desc"`.

In addition to the query parameters, the user can provide the maximum number of entities to be retrieved by using the parameter `maxentities` (`-m`), set to 10000 by default. Use -1 to retrieve all entities. Example: `-m 100` or `-m -1`.

Note that OpenAlex API recommends downloading and processing the snapshots of the dataset instead of using the API if you plan to download a large chunk of the complete dataset.

### [JSON Lines output](#json-lines-output)
The output can be saved to a JSON Lines file (each line containing a JSON entry) by passing the argument `--outputfile` (`-o`). Example: `-o works.jsonl`.

### [Aggregating queries](#aggregating-queries)
It is also possible to combine several queries by providing a `.csv` or `.tsv` file with the queries. The file should have the following columns: `filter`, `search`, `sort` and `maxentities`. Missing columns will be filled with the default values. The output will have all the aggregated queries. Example: `openalexnet -i queries.csv` for a file `queries.csv` with the following content:
```csv
filter,search,sort,maximum_entities
"type:journal-article","""complex networks""","cited_by_count:desc",10000
"type:journal-article","""network science""","cited_by_count:desc",10000
```
This should retrieve the 10000 most cited works with the terms "complex networks" or "network science" using two different queries. The folder `Examples/query_files/` provides more examples of query files.

### [Generating networks](#generating-networks)
The command-line application can also generate citation and coauthorship networks from the retrieved entities. The networks can be saved in 3 different formats: `.edgelist`, `.gml`, or `.xnet`.
The citation network can be generated by providing the argument `--citationfile` (`-c`), with the parameter being the file path where the network should be saved. The extension of the file will determine the format. Example: `-c citation_network.gml`. Similarly, the coauthorship network can be generated by providing the argument `--coauthorfile` (`-a`). Example: `-c citation_network.gml -a coauthorship_network.gml`.

Attributes of works can be selected to be exported in the network by providing the argument `--keptattributes` (`-k`). The attributes should be comma-separated. Example: `-n "id,title,doi"`.

By default the following properties are exported in the network:
```
id, doi, title, display_name, publication_year, publication_date, type, authorships, concepts, host_venue
```

The parameter --ignoreattributes (`-g`) can be used to ignore some of the default attributes. Example: `-i "authorships,concepts,host_venue"`.

For the case of coauthorship networks, the user can provide two extra parameters:
 - `--no_simplenetworks` (`-n`): If enabled, the coauthorship network edges will not be aggregated, resulting in multiple edges. The default is disabled.
 - `--countweights` (`-w`) If enabled the coauthorship network will have non-normalized weights, i.e., the contribution of a paper to a connection weight is 1.0, otherwise the contribution is the inverse of the number of authors in the paper. The default is disabled.

if `.edgelist` format is used, extra `csv` files with the nodes and edges attributes will be generated with the same name as the network file, but with the extension `_nodes.csv` and `_edges.csv`.

### [Loading from saved JSON Lines files](#loading-from-saved-json-lines-files)
The command-line application can also load the JSON Lines files generated by the API and generate the networks. This can be done by providing the argument `--inputfile` (`-i`). Example: `-i works.jsonl -c citation_network.gml -a coauthorship_network.gml`.

### [Polite mode](#polite-mode)
Finally, users can use the polite mode by providing an email address using `--email` (`-e`). See https://docs.openalex.org/how-to-use-the-api/ for more information.

### [Example usage](#example-usage)
To obtain the works with the term`"complex networks"` (in abstracts, titles or fulltexts) sorted by the number of citations. This also generates gml files for the citation and coauthorship networks.
```bash
openalexnet -t works -f "type:journal-article" -s "complex networks" -r "cited_by_count:desc" -o works.jsonl -c citation_network.gml -a coauthorship_network.gml
```
Note that because `maxentities` is not provided, only the 10000 most cited works will be obtained.

To load the saved works.jsonl file and generate the networks:
```bash
openalexnet -t works -i works.jsonl -c citation_network.edgelist -a coauthorship_network.edgelist
```

Use a query file to retrieve works and save them to a JSON Lines file:
```bash
openalexnet -t works -q query.csv -o works.jsonl
```

## [Python Library Usage](#python-library-usage)

Obtaining works from a specific author:

```python
    filterData = {
        "author.id": "A2420755856", # Eugene H. Stanley
        "is_paratext": "false",  # Only works, no paratexts (https://en.wikipedia.org/wiki/Paratext)
        "type": "journal-article", # Only journal articles
        "from_publication_date": "2000-01-01" # Published after 2000
    }

entityType = "works"

openalex = oanet.OpenAlexAPI() # add your email to accelerate the API calls. See https://openalex.org/api

entities = openalex.getEntities(entityType, filter=filterData)

entitiesList = []
    for entity in tqdm(entities,desc="Retrieving entries"):
        entitiesList.append(entity)

# Saving data as json lines (each line is a json object)
    oanet.saveJSONLines(entitiesList,"works_filtered.jsonl")
```

Check `Examples` folder for more examples.

## [Coming soon](#coming-soon)
 - Full API documentation
    - More examples
 - Unit tests
 - Group count

## Google Colaboratory Demo/Tutorial
You can access a Google Colab demo and tutorial by using the following link.
<a target="_blank" href="https://colab.research.google.com/github/filipinascimento/openalexnet">
  <img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"/>
</a>

## [Thanks](#thanks)
Remember to cite the OpenAlex work:
```bib
@article{priem2022openalex,
  title={OpenAlex: A fully-open index of scholarly works, authors, venues, institutions, and concepts},
  author={Priem, Jason and Piwowar, Heather and Orr, Richard},
  journal={arXiv preprint arXiv:2205.01833},
  year={2022}
}
```
If you use this code, please give it a star and share with your coleagues. Also stay tuned as I plan to develop a web-based interface for dynamic visualization of openalex networks. Check out [Helios-Web](http://github.com/filipinascimento/helios-web) to see the development progress of our network visualization tools.

openalexnet

Explore Similar Repositories

Was this content helpful?

Recent Feedback

Download README