diff --git a/app.py b/app.py index 4c0fba7..fc029d2 100644 --- a/app.py +++ b/app.py @@ -43,7 +43,7 @@ app_ui = ui.page_navbar( ), align="right", ), - selected="intro", + selected="search_engine", fluid=False, title=ui.div(ui.img(src="favicon.ico", width="75dpi", height="75dpi"), ui.h1("Copbird")), diff --git a/src/mod_searchable.py b/src/mod_searchable.py index 509f4be..bea73a2 100644 --- a/src/mod_searchable.py +++ b/src/mod_searchable.py @@ -4,6 +4,7 @@ from sklearn.metrics.pairwise import cosine_similarity import pandas as pd import numpy as np import pickle +import re tfidf_matrix_path = "data/tfidf_matrix.pckl" tfidf_vectorizer_path = "data/tfidf_vectorizer.pckl" @@ -13,6 +14,21 @@ tweets_path = "data/tweets_all_combined.csv" reply_html_svg = '' retweet_html_svg = '' like_html_svg = '' +quote_count_svg = ' ' + +link_regex = r"(https?://\S+)" +hashtag_regex = r"#(\w+)" + + +def replace_link(match): + url = match.group(0) + return f'{url}' + + +def replace_hastag(match): + hashtag = match.group(0) + name = str(hashtag).removeprefix("#") + return f'{hashtag}' print("Loading data from storage") @@ -29,10 +45,10 @@ with open(tfidf_vectorizer_path, "rb") as f: tweets["relevance_score"] = relevance_score["relevance_score"] -tweets = tweets.drop(["user_id", "measured_at", "tweet_id"], axis=1) +tweets = tweets.drop(["user_id", "measured_at"], axis=1) -def search_query(query: str, limit: int = 5) -> pd.DataFrame: +def search_query(query: str, limit: int = 5, sorting_method: str = "score") -> (pd.DataFrame, int): query_vec = tfidf_vectorizer.transform([query]) similarity = cosine_similarity(query_vec, tfidf_matrix).flatten() @@ -44,32 +60,60 @@ def search_query(query: str, limit: int = 5) -> pd.DataFrame: if not len(result): return None - overall = result['relevance_score'] * similarity[correct_indices] - return result.loc[overall.sort_values(ascending=False).index].head(limit) + if limit == -1: + limit = len(result) + + results = None + if sorting_method == "score": + overall = (0.6 * result['relevance_score']) * similarity[correct_indices] + results = result.loc[overall.sort_values(ascending=False).index].head(limit) + elif sorting_method == "date_new": + results = result.sort_values(by="created_at", ascending=False).head(limit) + elif sorting_method == "date_old": + results = result.sort_values(by="created_at", ascending=True).head(limit) + + return results, len(result) -@module.ui +@ module.ui def searchable_ui(): return ui.div( ui.h2("Tweet Suchmaschine"), - ui.input_text("search_input", "Suche:", placeholder="Gebe Suchterm ein", value="Leipzig"), - ui.HTML("
"), + ui.HTML("
"), + ui.row( + ui.column(6, ui.input_text("search_input", "Suche", placeholder="Gib Suchterm ein", value="Leipzig", width="100%")), + ui.column(3, + ui.input_select("sorting_method", "Sortierung", {"score": "Relevanz", "date_new": "Neuste Zuerst", "date_old": "Älteste Zuerst"}, selected="score", selectize=True, width="12em"), + ui.input_select("tweet_count", "Ergebnisse", {"5": "5", "20": "20", "50": "50", "all": "alle"}, selected="5", selectize=True, width="12em"), + style="display: flex; flex-direction: column; align-items: center; justify-content: center;"), + style="justify-content:space-between;" + + ), ui.output_ui(id="searchable_tweet_ui"), ) @ module.server def searchable_server(input: Inputs, output: Outputs, session: Session): - @output - @render.ui + @ output + @ render.ui def searchable_tweet_ui(): query = input.search_input() + sorting_method = input.sorting_method() + result_count = str(input.tweet_count()) + if result_count == "all": + result_count = -1 + else: + result_count = int(result_count) - result_pd = search_query(query, 15) + result_pd, found_tweets = search_query(query, result_count, sorting_method=sorting_method) - style = "text-align: center; padding-top: 0.5em;" - tweet_ui = ui.page_fluid() + style = "text-align: center;" + tweet_ui = ui.page_fluid( + ui.HTML(f"Gesamt gefundene Tweets: {found_tweets}"), + ui.HTML("

"), + ) if result_pd is None: return ui.div( @@ -78,27 +122,53 @@ def searchable_server(input: Inputs, output: Outputs, session: Session): # iterating over dataframe is bad but needed for idx, row in result_pd.iterrows(): + + # prettify tweet text + tweet_text = str(row["tweet_text"]).replace("\\n", "
") + tweet_text = re.sub(link_regex, replace_link, tweet_text) + tweet_text = re.sub(hashtag_regex, replace_hastag, tweet_text) + + tweet_link = f"https://twitter.com/{row['handle']}/status/{row['tweet_id']}" + + user_handle = row['handle'] + user_name = row['user_name'] + tweet_ui.append( ui.div( ui.row( - ui.column(9, ui.markdown( - f"**{row['user_name']}** *@{row['handle']}*"), style=style), - ui.column(3, ui.p(f"{row['created_at']}"), style=style), + ui.column(6, ui.HTML( + f"{user_name} @{user_handle}"), style=style + "padding-top: 1.5em; "), + ui.column(6, ui.p(f"{row['created_at']}"), style=style + "padding-top: 1.5em;"), ), ui.row( - ui.column(12, ui.HTML(str(row["tweet_text"]).replace( - "\\n", "
")), style=style + "font-size: 20px; padding:1em;"), + ui.column(12, ui.HTML("
"), + ui.HTML(f""" + +
{tweet_text}
+ """), + ui.HTML("
")), ), ui.row( ui.column(3, ui.HTML(reply_html_svg), ui.p( - f"{row['reply_count']}"), style=style), + f"{row['reply_count']}"), style=style, title="Antworten"), ui.column(3, ui.HTML(retweet_html_svg), ui.p( - f"{row['retweet_count']}"), style=style), + f"{row['retweet_count']}"), style=style, title="Retweets"), ui.column(3, ui.HTML(like_html_svg), ui.p( - f"{row['like_count']}"), style=style), - # quote_count: . Indicates approximately how many times this Tweet has been quoted by Twitter users. Example: - # TODO: use a nice svg for quote_count - ui.column(3, ui.p(f"Quote Count: {row['quote_count']}"), style=style), - ), style="border: 1px solid #954; margin-bottom: 1em;")) + f"{row['like_count']}"), style=style, title="Likes"), + ui.column(3, ui.HTML(quote_count_svg), ui.p( + f"{row['quote_count']}"), style=style, title="Quotes"), + ), style="border: 2px solid #119; margin-bottom: 1.5em; border-radius: 10px;")) return tweet_ui