naco-api/internal/twitter/client.go

package twitter

import (
	"context"
	"encoding/json"
	"fmt"
	"io"
	"net/http"
	"net/url"
	"strconv"
	"strings"
	"time"

	log "github.com/nacorid/logger"
)

func parseTwitterTime(s string) time.Time {
	t, err := time.Parse(time.RFC3339, s)
	if err != nil {
		log.Warn("Failed to parse Twitter timestamp", "timestamp", s, "error", err)
		return time.Time{}
	}
	return t
}

const apiBase = "https://api.x.com/2"

// httpTimeout bounds a single HTTP request so a stalled X backend cannot block
// a request indefinitely.
const httpTimeout = 60 * time.Second

const userAgent = "naco-api/1.0"

type Client struct {
	httpClient  *http.Client
	bearerToken string
}

func NewTwitterClient(bearerToken string) *Client {
	return &Client{
		httpClient:  &http.Client{Timeout: httpTimeout},
		bearerToken: bearerToken,
	}
}

type Post struct {
	ID        string
	Author    string
	Content   string
	CreatedAt time.Time
	Likes     int
	Reposts   int
	Quotes    int
	Replies   int
}

type userResponse struct {
	Data struct {
		ID       string `json:"id"`
		Name     string `json:"name"`
		Username string `json:"username"`
	} `json:"data"`
	Errors []struct {
		Title  string `json:"title"`
		Detail string `json:"detail"`
	} `json:"errors"`
}

type tweetMetrics struct {
	RetweetCount int `json:"retweet_count"`
	ReplyCount   int `json:"reply_count"`
	LikeCount    int `json:"like_count"`
	QuoteCount   int `json:"quote_count"`
}

type tweetData struct {
	ID            string       `json:"id"`
	Text          string       `json:"text"`
	CreatedAt     string       `json:"created_at"`
	PublicMetrics tweetMetrics `json:"public_metrics"`
}

type tweetsResponse struct {
	Data     []tweetData `json:"data"`
	Includes struct {
		Users []struct {
			Username string `json:"username"`
		} `json:"users"`
	} `json:"includes"`
	Errors []struct {
		Title  string `json:"title"`
		Detail string `json:"detail"`
	} `json:"errors"`
}

type singleTweetResponse struct {
	Data struct {
		ID            string       `json:"id"`
		Text          string       `json:"text"`
		CreatedAt     string       `json:"created_at"`
		PublicMetrics tweetMetrics `json:"public_metrics"`
	} `json:"data"`
	Includes struct {
		Users []struct {
			Username string `json:"username"`
		} `json:"users"`
	} `json:"includes"`
	Errors []struct {
		Title  string `json:"title"`
		Detail string `json:"detail"`
	} `json:"errors"`
}

const (
	maxRetries       = 4
	maxRateLimitWait = 30 * time.Second // Never stall a request longer than 30s
)

var defaultBackoff = 500 * time.Millisecond

func (c *Client) doGet(ctx context.Context, endpoint string) ([]byte, error) {
	backoff := defaultBackoff

	for attempt := 0; ; attempt++ {
		req, err := http.NewRequestWithContext(ctx, http.MethodGet, endpoint, nil)
		if err != nil {
			return nil, fmt.Errorf("failed to build request: %w", err)
		}
		req.Header.Set("Authorization", "Bearer "+c.bearerToken)
		req.Header.Set("User-Agent", userAgent)

		resp, err := c.httpClient.Do(req)
		if err != nil {
			// Transport-level failure (connection refused/reset, timeout).
			if attempt >= maxRetries {
				return nil, fmt.Errorf("twitter request failed after %d attempts: %w", maxRetries+1, err)
			}
			log.WarnWithContext(ctx, "twitter transport error, waiting to retry", "attempt", attempt+1, "error", err)
			if werr := waitForRetry(ctx, backoff); werr != nil {
				return nil, werr
			}
			backoff *= 2
			continue
		}

		// Handle 429 Too Many Requests
		if resp.StatusCode == http.StatusTooManyRequests {
			// Drain & close body before retrying.
			_, _ = io.Copy(io.Discard, resp.Body)
			_ = resp.Body.Close()

			if attempt >= maxRetries {
				return nil, fmt.Errorf("twitter rate limit exceeded, max retries reached")
			}

			wait := parseRateLimitWait(resp.Header, backoff)

			// Protect against 15-minute thread freezes
			if wait > maxRateLimitWait {
				return nil, fmt.Errorf("twitter rate limited: reset wait (%s) exceeds maximum allowed (%s)", wait, maxRateLimitWait)
			}

			log.WarnWithContext(ctx, "twitter rate limited, waiting to retry", "attempt", attempt+1, "wait", wait.String())
			if werr := waitForRetry(ctx, wait); werr != nil {
				return nil, werr
			}
			backoff *= 2
			continue
		}

		body, err := io.ReadAll(io.LimitReader(resp.Body, 2*1024*1024))
		_ = resp.Body.Close()
		if err != nil {
			return nil, fmt.Errorf("failed to read response: %w", err)
		}

		// Retry transient server-side failures (5xx) with backoff. 4xx/3xx fail
		// fast below.
		if resp.StatusCode >= http.StatusInternalServerError && attempt < maxRetries {
			log.WarnWithContext(ctx, "twitter server error, waiting to retry", "attempt", attempt+1, "status", resp.Status)
			if werr := waitForRetry(ctx, backoff); werr != nil {
				return nil, werr
			}
			backoff *= 2
			continue
		}

		if resp.StatusCode != http.StatusOK {
			errMsg := strings.TrimSpace(string(body))
			if len(errMsg) > 512 {
				errMsg = errMsg[:512] + "... (truncated)"
			}
			return nil, fmt.Errorf("twitter error (%s): %s", resp.Status, errMsg)
		}

		return body, nil
	}
}

// waitForRetry sleeps for wait, aborting early if the context is canceled.
// A timer is used (not time.After) so the timer can be stopped to free its
// resource when the caller abandons the wait.
func waitForRetry(ctx context.Context, wait time.Duration) error {
	timer := time.NewTimer(wait)
	defer timer.Stop()
	select {
	case <-timer.C:
		return nil
	case <-ctx.Done():
		return ctx.Err()
	}
}

// parseRateLimitWait extracts x-rate-limit-* or Retry-After, falling back to backoff.
func parseRateLimitWait(header http.Header, fallback time.Duration) time.Duration {
	// 1. Check x-rate-limit-reset (Unix timestamp)
	if remStr := header.Get("x-rate-limit-remaining"); remStr != "" {
		if rem, err := strconv.Atoi(remStr); err == nil && rem <= 0 {
			if rstStr := header.Get("x-rate-limit-reset"); rstStr != "" {
				if rst, err := strconv.ParseInt(rstStr, 10, 64); err == nil {
					// Add 1s clock skew buffer
					wait := time.Until(time.Unix(rst, 0)) + (1 * time.Second)
					if wait > 0 {
						return wait
					}
				}
			}
		}
	}

	// 2. Check standard Retry-After (seconds)
	if ra := header.Get("Retry-After"); ra != "" {
		if secs, err := strconv.Atoi(ra); err == nil && secs > 0 {
			return time.Duration(secs)*time.Second + (1 * time.Second)
		}
	}

	return fallback
}

func (c *Client) resolveUser(ctx context.Context, username string) (id, handle string, err error) {
	endpoint := fmt.Sprintf("%s/users/by/username/%s", apiBase, url.PathEscape(username))
	body, err := c.doGet(ctx, endpoint)
	if err != nil {
		return "", "", err
	}
	var ur userResponse
	if err := json.Unmarshal(body, &ur); err != nil {
		return "", "", fmt.Errorf("failed to unmarshal user response: %w", err)
	}
	if len(ur.Errors) > 0 {
		return "", "", fmt.Errorf("twitter error: %s: %s", ur.Errors[0].Title, ur.Errors[0].Detail)
	}
	if ur.Data.ID == "" {
		return "", "", fmt.Errorf("twitter user %q not found", username)
	}
	return ur.Data.ID, ur.Data.Username, nil
}

func (c *Client) GetLatestNPosts(ctx context.Context, actor string, limit int) ([]*Post, error) {
	if c.bearerToken == "" {
		return nil, fmt.Errorf("twitter bearer token not configured")
	}
	username := strings.TrimPrefix(actor, "@")

	userID, handle, err := c.resolveUser(ctx, username)
	if err != nil {
		return nil, err
	}

	if limit > 100 {
		limit = 100
	}

	endpoint := fmt.Sprintf("%s/users/%s/tweets?max_results=%d&tweet.fields=created_at,public_metrics&exclude=retweets,replies",
		apiBase, url.PathEscape(userID), limit)
	return c.fetchTweets(ctx, handle, endpoint)
}

func (c *Client) GetPostsInTimeframe(ctx context.Context, actor string, start, end time.Time, limit int) ([]*Post, error) {
	if c.bearerToken == "" {
		return nil, fmt.Errorf("twitter bearer token not configured")
	}
	username := strings.TrimPrefix(actor, "@")

	userID, handle, err := c.resolveUser(ctx, username)
	if err != nil {
		return nil, err
	}
	if limit > 100 {
		limit = 100
	}

	endpoint := fmt.Sprintf("%s/users/%s/tweets?start_time=%s&end_time=%s&max_results=%d&tweet.fields=created_at,public_metrics&exclude=retweets,replies",
		apiBase, url.PathEscape(userID), start.Format(time.RFC3339), end.Format(time.RFC3339), limit)
	return c.fetchTweets(ctx, handle, endpoint)
}

// fetchTweets performs the request and converts the tweet payload into Posts,
// skipping entries without an ID so downstream dedupe keys are never empty.
func (c *Client) fetchTweets(ctx context.Context, username, endpoint string) ([]*Post, error) {
	body, err := c.doGet(ctx, endpoint)
	if err != nil {
		return nil, err
	}

	var tr tweetsResponse
	if err := json.Unmarshal(body, &tr); err != nil {
		return nil, fmt.Errorf("failed to unmarshal tweets response: %w", err)
	}
	if len(tr.Errors) > 0 {
		return nil, fmt.Errorf("twitter error: %s: %s", tr.Errors[0].Title, tr.Errors[0].Detail)
	}

	posts := make([]*Post, 0, len(tr.Data))
	for _, t := range tr.Data {
		if t.ID == "" {
			continue
		}
		posts = append(posts, &Post{
			ID:        t.ID,
			Author:    username,
			Content:   t.Text,
			CreatedAt: parseTwitterTime(t.CreatedAt),
			Likes:     t.PublicMetrics.LikeCount,
			Reposts:   t.PublicMetrics.RetweetCount,
			Quotes:    t.PublicMetrics.QuoteCount,
			Replies:   t.PublicMetrics.ReplyCount,
		})
	}
	return posts, nil
}

func (c *Client) GetPost(ctx context.Context, id string) (*Post, error) {
	if c.bearerToken == "" {
		return nil, fmt.Errorf("twitter bearer token not configured")
	}
	endpoint := fmt.Sprintf("%s/tweets/%s?tweet.fields=created_at,public_metrics&expansions=author_id&user.fields=username",
		apiBase, url.PathEscape(id))
	body, err := c.doGet(ctx, endpoint)
	if err != nil {
		return nil, err
	}

	var tr singleTweetResponse
	if err := json.Unmarshal(body, &tr); err != nil {
		return nil, fmt.Errorf("failed to unmarshal tweet response: %w", err)
	}
	if len(tr.Errors) > 0 {
		return nil, fmt.Errorf("twitter error: %s: %s", tr.Errors[0].Title, tr.Errors[0].Detail)
	}
	if tr.Data.ID == "" {
		return nil, fmt.Errorf("tweet %q not found", id)
	}
	author := "unknown"
	if len(tr.Includes.Users) > 0 {
		author = tr.Includes.Users[0].Username
	}
	return &Post{
		ID:        tr.Data.ID,
		Author:    author,
		Content:   tr.Data.Text,
		CreatedAt: parseTwitterTime(tr.Data.CreatedAt),
		Likes:     tr.Data.PublicMetrics.LikeCount,
		Reposts:   tr.Data.PublicMetrics.RetweetCount,
		Quotes:    tr.Data.PublicMetrics.QuoteCount,
		Replies:   tr.Data.PublicMetrics.ReplyCount,
	}, nil
}