refactor: dynamic UA matching, session reuse, score Optional, --output on all commands

- constants.py: sync_chrome_version() aligns UA/sec-ch-ua with impersonate target - auth.py: reuse shared cffi session instead of creating duplicate - filter.py: eliminate double weight building in score_tweet - models.py: Tweet.score → Optional[float] for accurate display - cli.py: add --output to search/likes/user-posts for consistency
2026-03-09 21:15:28 +08:00
parent fda9b1c3dc
commit 8313a7012f
9 changed files with 79 additions and 41 deletions
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -4,7 +4,7 @@ build-backend = "hatchling.build"

 [project]
 name = "twitter-cli"
-version = "0.4.3"
+version = "0.4.4"
 description = "A CLI for Twitter/X — feed, bookmarks, and user timeline in terminal"
 readme = "README.md"
 license = "Apache-2.0"
--- a/twitter_cli/auth.py
+++ b/twitter_cli/auth.py
@@ -17,9 +17,7 @@ import subprocess
 import sys
 from typing import Dict, Optional

-from curl_cffi import requests as _cffi_requests
-
-from .constants import BEARER_TOKEN, USER_AGENT
+from .constants import BEARER_TOKEN, get_user_agent

 logger = logging.getLogger(__name__)

@@ -49,7 +47,7 @@ def verify_cookies(auth_token, ct0, cookie_string=None):
    Tries multiple endpoints. Only raises on clear auth failures (401/403).
    For other errors (404, network), returns empty dict (proceed without verification).
    """
-    from .client import _best_chrome_target
+    from .client import _get_cffi_session

    urls = [
        "https://api.x.com/1.1/account/verify_credentials.json",
@@ -65,14 +63,11 @@ def verify_cookies(auth_token, ct0, cookie_string=None):
        "X-Csrf-Token": ct0,
        "X-Twitter-Active-User": "yes",
        "X-Twitter-Auth-Type": "OAuth2Session",
-        "User-Agent": USER_AGENT,
+        "User-Agent": get_user_agent(),
    }

-    proxy = os.environ.get("TWITTER_PROXY", "")
-    session = _cffi_requests.Session(
-        impersonate=_best_chrome_target(),
-        proxies={"https": proxy, "http": proxy} if proxy else None,
-    )
+    # Reuse the shared curl_cffi session for consistent TLS fingerprint
+    session = _get_cffi_session()

    for url in urls:
        try:
--- a/twitter_cli/cli.py
+++ b/twitter_cli/cli.py
@@ -244,8 +244,9 @@ def user(screen_name):
@click.argument("screen_name")
@click.option("--max", "-n", "max_count", type=int, default=20, help="Max number of tweets to fetch.")
@click.option("--json", "as_json", is_flag=True, help="Output as JSON.")
-def user_posts(screen_name, max_count, as_json):
-    # type: (str, int, bool) -> None
+@click.option("--output", "-o", "output_file", type=str, default=None, help="Save tweets to JSON file.")
+def user_posts(screen_name, max_count, as_json, output_file):
+    # type: (str, int, bool, Optional[str]) -> None
    """List a user's tweets. SCREEN_NAME is the @handle (without @)."""
    screen_name = screen_name.lstrip("@")
    config = load_config()
@@ -258,7 +259,7 @@ def user_posts(screen_name, max_count, as_json):
        sys.exit(1)
    _fetch_and_display(
        lambda count: client.fetch_user_tweets(profile.id, count),
-        "@%s tweets" % screen_name, "📝", max_count, as_json, None, False, config,
+        "@%s tweets" % screen_name, "📝", max_count, as_json, output_file, False, config,
    )


@@ -277,15 +278,16 @@ SEARCH_PRODUCTS = ["Top", "Latest", "Photos", "Videos"]
 )
@click.option("--max", "-n", "max_count", type=int, default=20, help="Max number of tweets to fetch.")
@click.option("--json", "as_json", is_flag=True, help="Output as JSON.")
+@click.option("--output", "-o", "output_file", type=str, default=None, help="Save tweets to JSON file.")
@click.option("--filter", "do_filter", is_flag=True, help="Enable score-based filtering.")
-def search(query, product, max_count, as_json, do_filter):
-    # type: (str, str, int, bool, bool) -> None
+def search(query, product, max_count, as_json, output_file, do_filter):
+    # type: (str, str, int, bool, Optional[str], bool) -> None
    """Search tweets by QUERY string."""
    config = load_config()
    client = _get_client(config)
    _fetch_and_display(
        lambda count: client.fetch_search(query, count, product),
-        "'%s' (%s)" % (query, product), "🔍", max_count, as_json, None, do_filter, config,
+        "'%s' (%s)" % (query, product), "🔍", max_count, as_json, output_file, do_filter, config,
    )


@@ -293,9 +295,10 @@ def search(query, product, max_count, as_json, do_filter):
@click.argument("screen_name")
@click.option("--max", "-n", "max_count", type=int, default=20, help="Max number of tweets to fetch.")
@click.option("--json", "as_json", is_flag=True, help="Output as JSON.")
+@click.option("--output", "-o", "output_file", type=str, default=None, help="Save tweets to JSON file.")
@click.option("--filter", "do_filter", is_flag=True, help="Enable score-based filtering.")
-def likes(screen_name, max_count, as_json, do_filter):
-    # type: (str, int, bool, bool) -> None
+def likes(screen_name, max_count, as_json, output_file, do_filter):
+    # type: (str, int, bool, Optional[str], bool) -> None
    """Show tweets liked by a user. SCREEN_NAME is the @handle (without @)."""
    screen_name = screen_name.lstrip("@")
    config = load_config()
@@ -308,7 +311,7 @@ def likes(screen_name, max_count, as_json, do_filter):
        sys.exit(1)
    _fetch_and_display(
        lambda count: client.fetch_user_likes(profile.id, count),
-        "@%s likes" % screen_name, "❤️", max_count, as_json, None, do_filter, config,
+        "@%s likes" % screen_name, "❤️", max_count, as_json, output_file, do_filter, config,
    )


--- a/twitter_cli/client.py
+++ b/twitter_cli/client.py
@@ -15,7 +15,10 @@ from curl_cffi import requests as _cffi_requests
 from x_client_transaction import ClientTransaction
 from x_client_transaction.utils import generate_headers as _gen_ct_headers, get_ondemand_file_url

-from .constants import BEARER_TOKEN, USER_AGENT, SEC_CH_UA, SEC_CH_UA_MOBILE, SEC_CH_UA_PLATFORM
+from .constants import (
+    BEARER_TOKEN, SEC_CH_UA_MOBILE, SEC_CH_UA_PLATFORM,
+    get_sec_ch_ua, get_user_agent, sync_chrome_version,
+)
 from .models import Author, Metrics, Tweet, TweetMedia, UserProfile

 logger = logging.getLogger(__name__)
@@ -124,6 +127,7 @@ def _get_cffi_session():
        import os
        proxy = os.environ.get("TWITTER_PROXY", "")
        target = _best_chrome_target()
+        sync_chrome_version(target)  # align UA/sec-ch-ua with impersonate target
        _cffi_session = _cffi_requests.Session(
            impersonate=target,
            proxies={"https": proxy, "http": proxy} if proxy else None,
@@ -174,7 +178,7 @@ def _scan_bundles():
    _bundles_scanned = True

    try:
-        html = _url_fetch("https://x.com", {"user-agent": USER_AGENT})
+        html = _url_fetch("https://x.com", {"user-agent": get_user_agent()})
        script_pattern = re.compile(
            r'(?:src|href)=["\']'
            r'(https://abs\.twimg\.com/responsive-web/client-web[^"\']+\.js)'
@@ -685,12 +689,12 @@ class TwitterClient:
            "X-Twitter-Active-User": "yes",
            "X-Twitter-Auth-Type": "OAuth2Session",
            "X-Twitter-Client-Language": "en",
-            "User-Agent": USER_AGENT,
+            "User-Agent": get_user_agent(),
            "Origin": "https://x.com",
            "Referer": "https://x.com",
            "Accept": "*/*",
            "Accept-Language": "en-US,en;q=0.9",
-            "sec-ch-ua": SEC_CH_UA,
+            "sec-ch-ua": get_sec_ch_ua(),
            "sec-ch-ua-mobile": SEC_CH_UA_MOBILE,
            "sec-ch-ua-platform": SEC_CH_UA_PLATFORM,
            "Sec-Fetch-Dest": "empty",
--- a/twitter_cli/constants.py
+++ b/twitter_cli/constants.py
@@ -1,17 +1,50 @@
 """Shared constants for twitter-cli."""

+import re
+
 BEARER_TOKEN = (
    "AAAAAAAAAAAAAAAAAAAAANRILgAAAAAAnNwIzUejRCOuH5E6I8xnZz4puTs"
    "%3D1Zv7ttfk8LF81IUq16cHjhLTvJu4FA33AGWWjCpTnA"
 )

-USER_AGENT = (
-    "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) "
-    "AppleWebKit/537.36 (KHTML, like Gecko) "
-    "Chrome/133.0.0.0 Safari/537.36"
-)
+# Default Chrome version — updated by _best_chrome_target() at runtime
+_DEFAULT_CHROME_VERSION = "133"
+_chrome_version = _DEFAULT_CHROME_VERSION  # mutable, set by sync_chrome_version()

-# Chrome Client Hints — sent by modern Chrome on every request
-SEC_CH_UA = '"Chromium";v="133", "Not(A:Brand";v="99", "Google Chrome";v="133"'
+
+def sync_chrome_version(impersonate_target):
+    # type: (str) -> None
+    """Sync USER_AGENT / SEC_CH_UA with the actual impersonate target.
+
+    Called once when _get_cffi_session() picks a target (e.g. "chrome136").
+    """
+    global _chrome_version
+    match = re.search(r"(\d+)", impersonate_target)
+    if match:
+        _chrome_version = match.group(1)
+
+
+def get_user_agent():
+    # type: () -> str
+    return (
+        "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) "
+        "AppleWebKit/537.36 (KHTML, like Gecko) "
+        "Chrome/%s.0.0.0 Safari/537.36" % _chrome_version
+    )
+
+
+def get_sec_ch_ua():
+    # type: () -> str
+    return '"Chromium";v="%s", "Not(A:Brand";v="99", "Google Chrome";v="%s"' % (
+        _chrome_version, _chrome_version,
+    )
+
+
+# Static Client Hints
 SEC_CH_UA_MOBILE = "?0"
 SEC_CH_UA_PLATFORM = '"macOS"'
+
+# Legacy aliases — modules that import these get the default value.
+# _build_headers() should use get_user_agent() / get_sec_ch_ua() instead.
+USER_AGENT = get_user_agent()
+SEC_CH_UA = get_sec_ch_ua()
--- a/twitter_cli/filter.py
+++ b/twitter_cli/filter.py
@@ -21,7 +21,7 @@ DEFAULT_WEIGHTS = {


 def score_tweet(tweet, weights=None):
-    # type: (Tweet, Optional[FilterWeights]) -> float
+    # type: (Tweet, Optional[Dict[str, float]]) -> float
    """Calculate engagement score for a single tweet.

    Formula:
@@ -30,15 +30,18 @@ def score_tweet(tweet, weights=None):
            + w_replies × replies
            + w_bookmarks × bookmarks
            + w_views_log × log10(views)
+
+    Args:
+        weights: Pre-built weight dict. If None, uses DEFAULT_WEIGHTS.
    """
-    weight_map = _build_weights(weights or {})
+    w = weights or DEFAULT_WEIGHTS
    m = tweet.metrics
    return (
-        weight_map["likes"] * m.likes
-        + weight_map["retweets"] * m.retweets
-        + weight_map["replies"] * m.replies
-        + weight_map["bookmarks"] * m.bookmarks
-        + weight_map["views_log"] * math.log10(max(m.views, 1))
+        w.get("likes", 1.0) * m.likes
+        + w.get("retweets", 3.0) * m.retweets
+        + w.get("replies", 2.0) * m.replies
+        + w.get("bookmarks", 5.0) * m.bookmarks
+        + w.get("views_log", 0.5) * math.log10(max(m.views, 1))
    )


--- a/twitter_cli/models.py
+++ b/twitter_cli/models.py
@@ -49,7 +49,7 @@ class Tweet:
    lang: str = ""
    retweeted_by: Optional[str] = None
    quoted_tweet: Optional[Tweet] = None
-    score: float = 0.0
+    score: Optional[float] = None


@dataclass
--- a/twitter_cli/serialization.py
+++ b/twitter_cli/serialization.py
@@ -112,7 +112,7 @@ def tweet_from_dict(data: Dict[str, Any]) -> Tweet:
        lang=str(data.get("lang") or ""),
        retweeted_by=_optional_str(data.get("retweetedBy")),
        quoted_tweet=quoted_tweet,
-        score=float(data.get("score") or 0.0),
+        score=float(data["score"]) if data.get("score") is not None else None,
    )


--- a/uv.lock
+++ b/uv.lock
@@ -950,7 +950,7 @@ wheels = [

 [[package]]
 name = "twitter-cli"
-version = "0.4.2"
+version = "0.4.3"
 source = { editable = "." }
 dependencies = [
    { name = "beautifulsoup4" },