# get you Scholar id from your page
# (e.g., https://scholar.google.com/citations?user=73-D2jgAAAAJ)
author_id = "73-D2jgAAAAJ"

# register on https://serpapi.com/ and get your SerpApi key
os.environ["SERP_API_KEY"] = "<YOUR_KEY>"

# cache your results
out_path = Path("results")

# depending on the number of your citations, scraping might:
# 1. require a paid serpapi plan
# 2. take a while (~ 7 sec/citation)...
results = scrape_author(author_id, out_path=out_path)

# load results as pandas dataframe
author_path = out_path / author_id
results = load_yaml(author_path / "results.yaml")
df = get_citation_df(results, keep_warnings=True)
df.sort_values("article").head(3)

# authors that most often cited your articles
cols = ["name", "cited_by_name", "affil_name", "website"]
author_df = drop_and_count_duplicates(
    df, ["name", "author_id"], col_name="cited_by_name"
)
author_df.loc[:, cols].sort_values("cited_by_name", ascending=False).head(5)

# citing authors with most citations
cols = ["name", "citations", "affil_name", "website"]
author_df.loc[:, cols].sort_values("citations", ascending=False).head(10)

# map of citing authors
get_map(author_df, out_path=author_path, radius_scale=5, radius_log_base=2)

# affiliations that most often cited your articles
cols = ["affil_name", "cited_by_affil", "affil_country_name"]
affil_df = drop_and_count_duplicates(df, ["affil_name"], col_name="cited_by_affil")
affil_df.loc[:, cols].sort_values("cited_by_affil", ascending=False).head(5)

# affiliations citing specific articles
cols = ["citations", "affil_name", "affil_country_name", "cited_by_affil"]
for article in df.loc[:, "article"].unique():
    print(article)
    article_df = df.loc[df.loc[:, "article"] == article].dropna(subset=["affil_name"])
    article_df = article_df.sort_values("citations", ascending=False)
    article_df = drop_and_count_duplicates(
        article_df, ["affil_name"], col_name="cited_by_affil"
    )
    display(article_df.loc[:, cols].head(5))