repository·master·Indexed 19 days ago
https://github.com/danielnsilva/semanticscholarAn unofficial Python client library for interacting with Semantic Scholar APIs, including the Academic Graph, Recommendations, and Datasets APIs. It provides typed responses, asynchronous support via AsyncSemanticScholar, and streamlined navigation for paginated results. Key features include paper and author retrieval, keyword searching with bulk retrieval options, and automatic retry mechanisms for rate limiting.
The semanticscholar library provides several key capabilities for interacting with Semantic Scholar:
When performing queries that return multiple results, the semanticscholar library returns a PaginatedResults object. This object allows you to iterate through all pages of results seamlessly. You can treat the PaginatedResults instance as an iterable to access individual items across all pages without manually managing page offsets or tokens.
# Example pattern for iterating through paginated results
for result in paginated_results:
print(result)The library provides typed response objects for easy data extraction. You can access specific fields as attributes (e.g., paper.title).
If you need the original API response in its dictionary format, use the raw_data attribute. To see all available fields in the response object, use the keys() method.
paper = sch.get_paper('10.1093/mind/lix.236.433')
# Access typed attributes
print(paper.title)
# Access original JSON as a dictionary
print(paper.raw_data)
# List all available fields
print(paper.keys())The Datasets API allows you to access large snapshots of Semantic Scholar data. It is built around three core concepts:
'2023-12-01'). You can also use 'latest' to refer to the most recent release.'papers', 'authors', or 'publications'.Note: While this library can retrieve diffs, it does not support the actual process of applying those diffs to update a local dataset. For that, you should refer to official documentation regarding tools like Spark.
To use the library, import the SemanticScholar class, instantiate the client, and use its methods to fetch data. The client provides typed responses and streamlined navigation for paginated results.
# First, import the client from semanticscholar module
from semanticscholar import SemanticScholar
# You'll need an instance of the client to request data from the API
sch = SemanticScholar()
# Get a paper by its ID
paper = sch.get_paper('10.1093/mind/lix.236.433')
# Print the paper title
print(paper.title)To use the library, import the SemanticScholar class, instantiate a client, and use its methods to fetch data. The client provides typed responses and supports various Semantic Scholar APIs including Academic Graph, Recommendations, and Datasets.
# First, import the client from semanticscholar module
from semanticscholar import SemanticScholar
# You'll need an instance of the client to request data from the API
sch = SemanticScholar()
# Get a paper by its ID
paper = sch.get_paper('10.1093/mind/lix.236.433')
# Print the paper title
print(paper.title)To install the most recent stable version of the semanticscholar library, use the standard pip install command.
pip install semanticscholarMethods like search_paper and search_author return paginated results (default 100 items per page).
[item for item in results]) automatically handles pagination and retrieves all available items..items property to access only the items in the current batch without triggering extra API calls..next_page() to fetch the next batch of results and append them to the current list.# Iterate through all results automatically
results = sch.search_paper('Computing Machinery and Intelligence')
all_results = [item for item in results]
# Access only the first page
first_page = sch.search_paper('Computing Machinery and Intelligence').items
# Fetch the next page manually
results = sch.search_paper('Computing Machinery and Intelligence')
results.next_page()
first_two_pages = results.itemsIf you need the latest features from the development branch, you can install directly from GitHub using one of two methods:
Method 1: Manual Clone Clone the repository locally and install the current directory.
Method 2: VCS Support Install directly via a Git URL using pip.
# Method 1: Manual Clone
git clone git@github.com:danielnsilva/semanticscholar.git
cd semanticscholar
pip install .
# Method 2: VCS Support
pip install git+https://github.com/danielnsilva/semanticscholar@masterTo use an API key for authenticated requests, pass the api_key argument when initializing the SemanticScholar class.
from semanticscholar import SemanticScholar
sch = SemanticScholar(api_key='your_api_key_here')Install the unofficial Python client for Semantic Scholar APIs using pip.
pip install semanticscholarTo use the library, initialize the SemanticScholar class and call its methods (such as get_paper) to retrieve data. The returned objects allow you to access attributes directly.
from semanticscholar import SemanticScholar
sch = SemanticScholar()
paper = sch.get_paper('10.1093/mind/lix.236.433')
print(paper.title)