from fasthtml.common import *
from datetime import datetime, timedelta
import requests
from datetime import datetime
import json
from markdown import markdown
from dotenv import load_dotenv
loaded = load_dotenv("./.env", override=True)
print("Loaded .env file:", loaded)
API_URL = os.getenv("API_URL")
API_KEY = os.getenv("MS_SEARCH_KEY")
css_content = open("styles.css").read()
app, rt = fast_app(hdrs=(Style(css_content),))
md_exts='codehilite', 'smarty', 'extra', 'sane_lists'
def Markdown(s, exts=md_exts, **kw): return Div(NotStr(markdown(s, extensions=exts)), **kw)
def date_range_inputs(start_date, end_date):
return Div(
Input(type="date", name="start_date", value=start_date.strftime("%Y-%m-%d"), title="Start date"),
Input(type="date", name="end_date", value=end_date.strftime("%Y-%m-%d"), title="End date"),
cls="date-range"
)
def search_form(start_date, end_date):
return Form(
Input(type="text", name="query", placeholder="Enter search query"),
date_range_inputs(start_date, end_date),
Button("Search", type="submit"),
hx_post="/search",
hx_target="#search-results",
hx_trigger="submit"
)
def iso_to_unix_timestamp(iso_string):
dt = datetime.fromisoformat(iso_string)
return int(dt.timestamp())
def unix_timestamp_to_nice_format(timestamp):
dt = datetime.fromtimestamp(timestamp)
return dt.strftime("%b %d, %Y")
def make_query(query, start_date, end_date, page=1, limit=10):
url = f"{API_URL}/indexes/comments/search"
headers = {
"Content-Type": "application/json",
"Authorization": f"Bearer {API_KEY}",
}
after_timestamp = iso_to_unix_timestamp(start_date)
before_timestamp = iso_to_unix_timestamp(end_date)
query = {
"q": query,
"limit": limit,
"offset": (page - 1) * limit,
"filter": f"comment_updatedAt_timestamp >= {after_timestamp} AND comment_updatedAt_timestamp < {before_timestamp}",
"attributesToCrop": ['comment_text'],
"cropLength": 30,
"attributesToHighlight": ["comment_text", "discussion_title"],
"highlightPreTag": "",
"highlightPostTag": ""
}
response = requests.post(url, headers=headers, json=query)
return response.json()
def search_results(query, start_date, end_date, page=1):
raw_results = make_query(query, start_date, end_date, page)
return Div(
make_results_bar(raw_results),
Div(*[make_card(r) for r in raw_results["hits"]]),
make_pagination(query, start_date, end_date, page, raw_results["estimatedTotalHits"]),
id="search-results"
)
def make_results_bar(results):
processing_time = results["processingTimeMs"]
estimated_hits = results["estimatedTotalHits"]
return Div(
Div(f"Processing time: {processing_time}ms"),
Div(f"Estimated total hits: {estimated_hits}"),
cls="results-bar"
)
def make_card(result):
result = result["_formatted"]
url = f"https://hf.co/{result['repo_id']}/discussions/{result['discussion_num']}"
date = unix_timestamp_to_nice_format(int(result["comment_updatedAt_timestamp"]))
return Div(
Div(
Strong(NotStr(result["discussion_title"])),
P(NotStr(result["comment_text"]), cls="comment-text"),
Div(Span(date)),
A(url, href=url, target="_blank"),
),
cls="card-item"
)
def make_pagination(query, start_date, end_date, current_page, total_hits, limit=10):
total_pages = -(-total_hits // limit) # Ceiling division
children = []
if current_page > 1:
children.append(
Button("Previous",
hx_post=f"/search?page={current_page-1}",
hx_target="#search-results",
hx_include="[name='query'], [name='start_date'], [name='end_date']")
)
children.append(Span(f"Page {current_page} of {total_pages}"))
if current_page < total_pages:
children.append(
Button("Next",
hx_post=f"/search?page={current_page+1}",
hx_target="#search-results",
hx_include="[name='query'], [name='start_date'], [name='end_date']",
)
)
return Div(*children, cls="pagination")
@rt("/")
def get():
end_date = datetime.now()
start_date = end_date - timedelta(days=7)
return Titled(
"HF Discussion Search",
Div(
search_form(start_date, end_date),
Div(id="search-results"),
cls="container"
)
)
@rt("/search")
def post(query: str, start_date: str, end_date: str, page: int = 1):
return search_results(query, start_date, end_date, page)
serve()