Author SHA1 Message Date
jimandClaude Opus 5 d6637266b7 fix: exclude inactive blogs from blog selection queries
Blogs.IsActive was honored only by GetBlogs and GetBlogsAll, so a blog
with IsActive = 0 was still selected for likes crawling and for output
mode. Add the filter to every remaining query that selects blog records:

- GetBlogsForLikes, all three variants (specific blog, ignoreCooldown,
  cooldown) - this is the selector that spends API quota
- GetAllBlogsWithTTFolderPath

Writes are deliberately untouched. The UPDATE statements are keyed on a
blog the caller already selected; filtering them would let the crawler
fetch a blog, pay the API cost, then fail to persist its cursor and
re-fetch the same pages on every run. Exclusion belongs at selection.

LegacyPostsDbImporter is also untouched: it reads a foreign legacy
schema that may not have the column.

Co-Authored-By: Claude Opus 5 <[email protected]>
2026-07-29 09:51:24 -05:00
jimandClaude Opus 4.8 2a02811003 Merge branch 'claude/rate-limit-behavior-474ecf'
Strip only a trailing numeric suffix from blog folder names, fixing
zomb-eh_10 importing as blog 'zomb-eh0'.

Co-Authored-By: Claude Opus 4.8 <[email protected]>
2026-07-22 16:40:16 -05:00
jimandClaude Opus 4.8 a73b597381 fix: strip only trailing numeric suffix from blog folder names
NormalizeBlogFolderName removed "_1".."_9" as unanchored substrings, so a
folder suffixed past a single digit lost the wrong characters: "_10" hit
the "_1" rule and left the trailing "0" welded to the name, importing
zomb-eh_10 as blog "zomb-eh0". That name does not exist on Tumblr, so
every post imported under it 404s on --collect forever.

Anchor the strip to a trailing _<digits> instead. This also fixes blogs
whose real name contains "_1" (some_1blog no longer becomes someblog) and
folders suffixed "_0", which were not stripped at all.

Verified against the live folder tree: zomb-eh_10 is the only existing
folder whose normalized name changes.

Co-Authored-By: Claude Opus 4.8 <[email protected]>
2026-07-22 12:58:35 -05:00
jim f9e1d2100b Merge remote master into local master 2026-07-22 12:12:33 -05:00
jimandClaude Opus 4.8 3e2b287737 Merge branch 'claude/rate-limit-behavior-474ecf'
Retry transient CDN failures instead of failing the post; throttle
--collect and --likes to 60/min.

Co-Authored-By: Claude Opus 4.8 <[email protected]>
2026-07-22 12:10:54 -05:00
jimandClaude Opus 4.8 003a504d5e fix: retry transient CDN failures instead of failing the post
A non-JSON response body (CDN 403/5xx HTML, empty body, transport error)
never reached the Tumblr API, so it says nothing about the post being
fetched. These were recorded as FAILURE, which consumed the post's single
attempt for the pass and cleared the API key's rate-limit flag on the way
through.

Classify them as Root.transientFailure and retry in place (1s/4s/10s)
before skipping. Skipped posts stay unmarked in the DB so a later launch
retries them. Ten consecutive transient failures now aborts the pass
rather than skipping post-by-post against an edge refusing all traffic.

Also:
- MarkAvailable() only on a response that reached the API, and it is now
  a no-op when the key was not flagged (was writing to the DB and logging
  on every single call)
- Only a real 429 counts as a rate limit; stop inferring one from
  X-RateLimit-* headers, which Tumblr sends on every response
- Limiters pace with AcquireAsync instead of AttemptAcquire, which did
  not wait and aborted the run once a window was saturated
- Throttle --collect and --likes from 300/min to 60/min
- Log one line per transient failure instead of the HTML body and stack
  trace; keep full detail only for a 2xx that fails to parse
- --collect returns exit 3 when a pass ends incomplete

Co-Authored-By: Claude Opus 4.8 <[email protected]>
2026-07-22 12:10:49 -05:00
jim 16147b273e docs: document default-mode .txt ingest field parsing in AGENTS.md
Note TraverseDirectory's recognized field prefixes, multi-line
Body/Downloaded files continuation, and how RootURL now gets
populated from both the .txt Reblog root url line and the API-based
--likes flow.
2026-07-16 10:43:16 -05:00
jim 5361bb78b8 Merge remote master into txt-validation branch 2026-07-16 10:42:28 -05:00
jim 21a5525094 Capture multi-line Body/Downloaded files in default .txt ingest mode
TraverseDirectory only ever read the single line immediately after
"Body:"/"Downloaded files:", silently dropping every continuation
line (multi-paragraph HTML bodies, multiple downloaded filenames).
Switch to an indexed line scan so those two fields collect lines
until the next recognized field prefix, matching how IngestMode.cs
already handles multi-line values.
2026-07-16 10:23:16 -05:00
jim 33839930e8 Parse Reblog root url in default .txt ingest mode
TraverseDirectory (the no-args ingest path) never read the "Reblog
root url:" line, so RootURL stayed unset even though AddPost/UpdatePost
already support it via the API-based --likes flow. New scraper output
now includes this field; wire it through both AddPost call sites.
2026-07-16 10:20:50 -05:00
4 changed files with 259 additions and 202 deletions
+20
View File
@@ -11,6 +11,7 @@
- `ResponseNotes.cs`: Tumblr API response models
- Round-robin API key rotation with rate-limit tracking
- Automatic console color assignment per API key for output differentiation
- No-argument mode (`Program.TraverseDirectory`) ingests `.txt` blog export files into `Posts` via `DataAccess.AddPost`. Recognized field prefixes live in `TraverseDirectoryFieldPrefixes`; `Body:` and `Downloaded files:` collect every following line up to the next recognized prefix (multi-line values). `RootURL` is populated from a `Reblog root url:` line the same way it's populated from the API-based `--likes` flow — both paths converge on `DataAccess.AddPost`'s `rootURL` parameter, which `UpdatePost` only overwrites when the incoming value is non-empty (existing `RootURL` is preserved otherwise)
## Developer Guidelines
@@ -27,6 +28,25 @@
- Preserve console color state: use save/restore pattern for temporary color changes
- API rate limits must use `ApiKeyPool.MarkRateLimited()`/`MarkAvailable()`
### API Failure Classification
Tumblr sits behind a CDN that returns HTML error pages (403, 5xx) which never reach the API. These
say nothing about the item being fetched, so they must not be recorded as per-item failures.
- A response body that will not parse as JSON did not come from the API. Flag it with
`Root.transientFailure`, never as `FAILURE`
- Transient failures retry in place (`TransientBackoffSeconds`) before the item is skipped; a skipped
item stays unmarked in the DB so a later launch retries it
- `MaxConsecutiveTransient` consecutive transient failures aborts the pass rather than skipping
item-by-item against an edge that is refusing all traffic
- Only call `ApiKeyPool.MarkAvailable()` on a response that actually reached the API. A transport or
CDN failure says nothing about the key's standing and must not clear its flag
- Only a real HTTP 429 (or `meta.status == 429`) counts as a rate limit. Do not infer one from the
presence of `X-RateLimit-*` headers, which Tumblr sends on every response
- Rate limiters must pace with `await AcquireAsync()`. `AttemptAcquire()` does not wait, so a
saturated window aborts the run instead of throttling it
- Long-running commands return exit 3 when a pass ends incomplete (rate-limit pause, breaker trip, or
skipped items), so a caller can distinguish that from a clean run
### Testing
- No existing test suite; use xUnit if adding tests
- Test critical logic: `ApiKeyPool` init, color parsing, config persistence
+62 -125
View File
@@ -37,6 +37,7 @@ namespace URLNotesGrabberCORE
public string reblogKey;
public string reblogName;
public string reblogURL;
public string rootURL;
public string slug;
public string summary;
public string tags;
@@ -60,6 +61,7 @@ namespace URLNotesGrabberCORE
reblogKey = ".";
reblogName = ".";
reblogURL = ".";
rootURL = ".";
slug = ".";
summary = ".";
tags = ".";
@@ -1038,7 +1040,8 @@ namespace URLNotesGrabberCORE
COALESCE(LikesCursor, 0),
COALESCE(LikesNewestTimestamp, 0)
FROM Blogs
WHERE BlogName = @blog";
WHERE BlogName = @blog
AND IsActive = 1";
}
else if (ignoreCooldown)
{
@@ -1051,6 +1054,7 @@ namespace URLNotesGrabberCORE
INNER JOIN Notes N ON N.NoteBlogName = B.BlogName
WHERE N.TimeStamp >= 1535778000
AND N.rootBlogName = B.BlogName
AND B.IsActive = 1
GROUP BY B.BlogName
ORDER BY MIN(N.Timestamp);";
}
@@ -1065,6 +1069,7 @@ namespace URLNotesGrabberCORE
INNER JOIN Notes N ON N.NoteBlogName = B.BlogName
WHERE N.TimeStamp >= 1535778000
AND N.rootBlogName = B.BlogName
AND B.IsActive = 1
AND (
B.LikesPulled = 0
OR COALESCE(B.LikesLastRefreshed, 0)
@@ -2209,7 +2214,7 @@ namespace URLNotesGrabberCORE
using var connection = new SQLiteConnection("Data Source=" + DBPath);
connection.Open();
using var cmd = new SQLiteCommand("SELECT BlogName, TTFolderPath FROM Blogs", connection);
using var cmd = new SQLiteCommand("SELECT BlogName, TTFolderPath FROM Blogs WHERE IsActive = 1", connection);
using var reader = cmd.ExecuteReader();
while (reader.Read())
{
@@ -2609,6 +2614,10 @@ namespace URLNotesGrabberCORE
public static void MarkAvailable(ApiKeyConfig key)
{
// Called after every successful call; skip the write and the log line when nothing was flagged.
if (GetRetryUntil(key) == 0)
return;
using var conn = new System.Data.SQLite.SQLiteConnection("Data Source=" + _dbPath);
conn.Open();
using var cmd = new System.Data.SQLite.SQLiteCommand(
@@ -2702,6 +2711,15 @@ namespace URLNotesGrabberCORE
private static string FormatKeyLabel(ApiKeyConfig key) => $"[Key#{key.KeyNumber}]";
private static string SummarizeBody(string body)
{
if (string.IsNullOrWhiteSpace(body))
return "(empty)";
var flat = System.Text.RegularExpressions.Regex.Replace(body, @"<[^>]+>|\s+", " ").Trim();
return flat.Length <= 80 ? flat : flat.Substring(0, 80) + "...";
}
private static int GetRetryDelaySecondsFromHeaders(IEnumerable<HeaderParameter>? headers)
{
if (headers == null)
@@ -2780,144 +2798,63 @@ namespace URLNotesGrabberCORE
Console.WriteLine($"{FormatKeyLabel(key)} {timestamp}\t{DateTime.Now}\t{DataAccess.UpdateAPICount()}");
var myDeserializedClass = new Root();
// Never reached the API: there is no body to interpret, so the post's state is still unknown.
if (response.ResponseStatus != ResponseStatus.Completed)
{
myDeserializedClass.statusCode = response.ResponseStatus.ToString();
myDeserializedClass.transientFailure = true;
Console.WriteLine($"[Transient] {FormatKeyLabel(key)} transport {response.ResponseStatus}: {response.ErrorException?.Message}");
return myDeserializedClass;
}
try
{
var deserializedResult = JsonConvert.DeserializeObject<Root>(myJsonResponse);
if (deserializedResult != null)
if (deserializedResult == null)
{
myDeserializedClass = deserializedResult;
myDeserializedClass.rawJson = myJsonResponse;
// Empty body behind an HTTP status: an edge/proxy response, not the API.
myDeserializedClass.statusCode = response.StatusCode.ToString();
myDeserializedClass.transientFailure = true;
Console.WriteLine($"[Transient] {FormatKeyLabel(key)} HTTP {(int)response.StatusCode} {response.StatusDescription} — empty body");
return myDeserializedClass;
}
if (myDeserializedClass.meta != null && myDeserializedClass.meta.status == 404)
{
myDeserializedClass.statusCode = "NotFound";
}
myDeserializedClass = deserializedResult;
myDeserializedClass.rawJson = myJsonResponse;
bool metaIndicatesRateLimit = myDeserializedClass.meta != null && myDeserializedClass.meta.status == 429;
bool metaMsgIndicatesRateLimit = myDeserializedClass.meta != null && !string.IsNullOrEmpty(myDeserializedClass.meta.msg) && myDeserializedClass.meta.msg.IndexOf("Too Many", StringComparison.OrdinalIgnoreCase) >= 0;
if (myDeserializedClass.meta != null && myDeserializedClass.meta.status == 404)
{
myDeserializedClass.statusCode = "NotFound";
}
if (metaIndicatesRateLimit || metaMsgIndicatesRateLimit || (response != null && (response.StatusDescription?.IndexOf("Too Many", StringComparison.OrdinalIgnoreCase) >= 0 || response.StatusCode == System.Net.HttpStatusCode.TooManyRequests)))
{
if (response?.Headers != null)
{
bool checkResetLocal = false;
foreach (var header in response.Headers)
{
string? headerName = header?.Name;
string? headerValue = header?.Value?.ToString();
if (string.IsNullOrEmpty(headerName) || string.IsNullOrEmpty(headerValue))
continue;
bool metaIndicatesRateLimit = myDeserializedClass.meta != null && myDeserializedClass.meta.status == 429;
bool metaMsgIndicatesRateLimit = myDeserializedClass.meta != null && !string.IsNullOrEmpty(myDeserializedClass.meta.msg) && myDeserializedClass.meta.msg.IndexOf("Too Many", StringComparison.OrdinalIgnoreCase) >= 0;
if (string.Equals(headerName, "Retry-After", StringComparison.OrdinalIgnoreCase))
{
if (int.TryParse(headerValue, out int retrySecs))
myDeserializedClass.retryInSeconds = Math.Max(myDeserializedClass.retryInSeconds, retrySecs);
else if (DateTimeOffset.TryParse(headerValue, out var dto))
myDeserializedClass.retryInSeconds = Math.Max(myDeserializedClass.retryInSeconds, (int)Math.Max(0, (dto - DateTimeOffset.UtcNow).TotalSeconds));
}
if (headerName.IndexOf("X-RateLimit-Reset", StringComparison.OrdinalIgnoreCase) >= 0 && long.TryParse(headerValue, out long epoch))
{
var secs = (int)Math.Max(0, epoch - DateTimeOffset.UtcNow.ToUnixTimeSeconds());
myDeserializedClass.retryInSeconds = Math.Max(myDeserializedClass.retryInSeconds, secs);
}
if (headerName.Contains("Remaining", StringComparison.OrdinalIgnoreCase) && headerValue == "0")
checkResetLocal = true;
if (checkResetLocal && headerName.IndexOf("Reset", StringComparison.OrdinalIgnoreCase) >= 0)
{
if (int.TryParse(headerValue, out int resetValue))
myDeserializedClass.retryInSeconds = Math.Max(myDeserializedClass.retryInSeconds, resetValue);
else if (long.TryParse(headerValue, out long epochVal))
myDeserializedClass.retryInSeconds = Math.Max(myDeserializedClass.retryInSeconds, (int)Math.Max(0, epochVal - DateTimeOffset.UtcNow.ToUnixTimeSeconds()));
}
}
}
myDeserializedClass.statusCode = "TooManyRequests";
}
if (metaIndicatesRateLimit || metaMsgIndicatesRateLimit || response.StatusDescription?.IndexOf("Too Many", StringComparison.OrdinalIgnoreCase) >= 0 || response.StatusCode == System.Net.HttpStatusCode.TooManyRequests)
{
myDeserializedClass.retryInSeconds = Math.Max(myDeserializedClass.retryInSeconds, GetRetryDelaySecondsFromHeaders(response.Headers));
myDeserializedClass.statusCode = "TooManyRequests";
}
}
catch (Exception ex)
{
Console.WriteLine($"Failed JSON: {myJsonResponse}");
Console.WriteLine(ex.ToString());
// A body that will not parse came from infrastructure (CDN/proxy/WAF), not the Tumblr
// API, so it says nothing about this post. Retryable, not a failure of the post itself.
myDeserializedClass.statusCode = response.StatusCode.ToString();
if (!response.IsSuccessful)
if (response.StatusCode == System.Net.HttpStatusCode.TooManyRequests)
{
string? statusStr = null;
try { statusStr = response != null ? response.StatusCode.ToString() : null; } catch { statusStr = null; }
Console.WriteLine($"{statusStr}\t{response?.StatusDescription}");
if (!string.IsNullOrEmpty(statusStr))
myDeserializedClass.statusCode = statusStr;
myDeserializedClass.retryInSeconds = GetRetryDelaySecondsFromHeaders(response.Headers);
myDeserializedClass.statusCode = "TooManyRequests";
}
else
{
myDeserializedClass.transientFailure = true;
Console.WriteLine($"[Transient] {FormatKeyLabel(key)} HTTP {(int)response.StatusCode} {response.StatusDescription} — unparseable body: {SummarizeBody(myJsonResponse)}");
bool checkReset = false;
if ((myDeserializedClass.statusCode != "NotFound" || myDeserializedClass.retryInSeconds > 0) && response.Headers != null)
{
bool foundRateLimitHeader = false;
foreach (var header in response.Headers)
{
if (header.Name != null && header.Value != null)
{
Console.WriteLine($"{header.Name} - {header.Value}");
var headerValue = header.Value?.ToString();
if (!string.IsNullOrEmpty(headerValue))
{
if (string.Equals(header.Name, "Retry-After", StringComparison.OrdinalIgnoreCase))
{
if (int.TryParse(headerValue, out int retrySecs))
{
if (myDeserializedClass.retryInSeconds < retrySecs)
myDeserializedClass.retryInSeconds = retrySecs;
}
else if (DateTimeOffset.TryParse(headerValue, out DateTimeOffset dto))
{
var secs = (int)Math.Max(0, (dto - DateTimeOffset.UtcNow).TotalSeconds);
if (myDeserializedClass.retryInSeconds < secs)
myDeserializedClass.retryInSeconds = secs;
}
}
if (checkReset && header.Name.Contains("Reset", StringComparison.OrdinalIgnoreCase))
{
if (int.TryParse(headerValue, out int resetValue))
{
if (myDeserializedClass.retryInSeconds < resetValue)
myDeserializedClass.retryInSeconds = resetValue;
}
else if (long.TryParse(headerValue, out long epochVal))
{
var secs = (int)Math.Max(0, epochVal - DateTimeOffset.UtcNow.ToUnixTimeSeconds());
if (myDeserializedClass.retryInSeconds < secs)
myDeserializedClass.retryInSeconds = secs;
}
}
if (header.Name.IndexOf("X-RateLimit-Reset", StringComparison.OrdinalIgnoreCase) >= 0)
{
if (long.TryParse(headerValue, out long epoch))
{
var secs = (int)Math.Max(0, epoch - DateTimeOffset.UtcNow.ToUnixTimeSeconds());
if (myDeserializedClass.retryInSeconds < secs)
myDeserializedClass.retryInSeconds = secs;
}
foundRateLimitHeader = true;
}
}
if (header.Name.Contains("Remaining", StringComparison.OrdinalIgnoreCase) && header.Value.ToString() == "0")
checkReset = true;
else
checkReset = false;
}
if (foundRateLimitHeader || (response != null && response.StatusCode == System.Net.HttpStatusCode.TooManyRequests))
{
myDeserializedClass.statusCode = "TooManyRequests";
}
}
}
// A 2xx that will not parse is a genuine surprise; keep the detail for that case only.
if (response.IsSuccessful)
Console.WriteLine(ex.ToString());
}
}
+173 -77
View File
@@ -281,7 +281,7 @@ namespace URLNotesGrabberCORE
managedCollectRun = true;
}
CollectNotes(settings.GetValue<string>("PathOutput"), withoutNotesOnly, beforeDate, managedCollectRun).GetAwaiter().GetResult();
exitCode = CollectNotes(settings.GetValue<string>("PathOutput"), withoutNotesOnly, beforeDate, managedCollectRun).GetAwaiter().GetResult();
break;
case "--blogsR": //collect notes from all posts
@@ -452,7 +452,7 @@ namespace URLNotesGrabberCORE
Console.WriteLine("--importposts [path-to-posts.db]\t One-time migration: copy legacy ThreeTxtFileHelper posts.db rows into TL.db");
Console.WriteLine();
Console.WriteLine("Exit status: 0 = success; 1 = unexpected error; 2 = usage error (unknown command or bad/missing arguments)");
Console.WriteLine("Exit status: 0 = success; 1 = unexpected error; 2 = usage error (unknown command or bad/missing arguments); 3 = incomplete (--collect paused on a rate limit, or skipped posts after transient API failures) - relaunch to resume");
}
static void WritePostBlogsToFile(string outPath)
@@ -582,16 +582,11 @@ namespace URLNotesGrabberCORE
protected static string NormalizeBlogFolderName(string folderName)
{
return folderName
.Replace("_1", "")
.Replace("_2", "")
.Replace("_3", "")
.Replace("_4", "")
.Replace("_5", "")
.Replace("_6", "")
.Replace("_7", "")
.Replace("_8", "")
.Replace("_9", "");
// Archive tools suffix duplicate blog folders with _1, _2, ... _10 and beyond. Strip only a
// trailing numeric suffix: unanchored substring removal ate the "_1" inside "_10" and left the
// "0" welded to the name (zomb-eh_10 -> zomb-eh0), and mangled any blog whose real name
// contains "_1". A blog name is never a prefix of itself plus "_<digits>", so this is safe.
return System.Text.RegularExpressions.Regex.Replace(folderName, @"_\d+$", "");
}
static async Task FetchAndStoreReplyText(string blogName, long postID, long timestamp)
@@ -847,7 +842,9 @@ namespace URLNotesGrabberCORE
RateLimiter limiter = new SlidingWindowRateLimiter(new SlidingWindowRateLimiterOptions
{
PermitLimit = 300,
// 1/sec average, matching --collect: the CDN reacts to aggregate traffic from the IP,
// not to per-command rates.
PermitLimit = 60,
QueueProcessingOrder = QueueProcessingOrder.OldestFirst,
QueueLimit = 1,
Window = TimeSpan.FromMinutes(1),
@@ -883,7 +880,9 @@ namespace URLNotesGrabberCORE
while (hasMoreLikes)
{
using RateLimitLease lease = limiter.AttemptAcquire(1);
// Wait for a permit rather than giving up on one: the limiter paces the loop, it is
// not a failure condition. Only one acquire is ever pending, so QueueLimit = 1 suffices.
using RateLimitLease lease = await limiter.AcquireAsync(1);
if (!lease.IsAcquired)
{
Console.WriteLine("!@@@@@@@ - Rate Limited Exceeded: No Lease Available");
@@ -1129,6 +1128,43 @@ if (shouldInsert)
}
}
// Backoff between in-place retries of a transient infrastructure failure. Most CDN 403s and edge
// 5xxs clear within a few seconds, so retrying here saves the post its single attempt for the pass.
static readonly int[] TransientBackoffSeconds = { 1, 4, 10 };
// Fetches one page, retrying transient failures in place. Rate limits are returned to the caller
// untouched — those are handled by pausing the whole run, not by retrying this post.
static async Task<Root> FetchNotesPage(Tuple<string, long, long, long> post, string beforeTimestamp)
{
Root response = null!;
for (int attempt = 0; ; attempt++)
{
var key = ApiKeyPool.GetCurrentKey();
response = await APIAccess.GrabNotes(key, post.Item1, post.Item2, beforeTimestamp);
if (response.statusCode == "TooManyRequests")
{
ApiKeyPool.MarkRateLimited(key, response.retryInSeconds > 0 ? response.retryInSeconds : 60);
return response;
}
if (!response.transientFailure)
{
// Only a response that actually reached the API says anything about the key's standing.
ApiKeyPool.MarkAvailable(key);
return response;
}
if (attempt >= TransientBackoffSeconds.Length)
return response;
int delay = TransientBackoffSeconds[attempt];
Console.WriteLine($"[Transient] retry {attempt + 1}/{TransientBackoffSeconds.Length} in {delay}s");
await Task.Delay(delay * 1000);
}
}
static async Task<string> GrabNotes(Tuple<string, long, long, long> post)
{
try
@@ -1146,18 +1182,16 @@ if (shouldInsert)
string beforeTimestamp = post.Item3.ToString();
bool hasReplies = false;
const int maxPages = 500;
var key = ApiKeyPool.GetCurrentKey();
var response = await APIAccess.GrabNotes(key, post.Item1, post.Item2, beforeTimestamp);
var response = await FetchNotesPage(post, beforeTimestamp);
if (response.statusCode == "TooManyRequests")
{
int retry = response.retryInSeconds > 0 ? response.retryInSeconds : 60;
ApiKeyPool.MarkRateLimited(key, retry);
return "TooManyRequests";
}
if (response.meta?.status != 429)
ApiKeyPool.MarkAvailable(key);
if (response.transientFailure)
{
Console.WriteLine($"[Skip] {post.Item1}/{post.Item2} — {response.statusCode} after {TransientBackoffSeconds.Length} retries");
return "Transient";
}
if (IsNotFound(response))
{
@@ -1166,19 +1200,6 @@ if (shouldInsert)
Thread.Sleep(1000);
return "NotFound";
}
if (response == null)
{
Console.WriteLine("##### Response is null - API Failure? ###");
return "FAILURE";
}
if (response.statusCode == "TooManyRequests")
{
int retry = response.retryInSeconds > 0 ? response.retryInSeconds : 60;
ApiKeyPool.MarkRateLimited(key, retry);
ApiKeyPool.SleepUntilAnyAvailable(30);
return response.statusCode;
}
// Pagination loop
while (true)
@@ -1228,16 +1249,15 @@ if (shouldInsert)
break;
}
key = ApiKeyPool.GetCurrentKey();
response = await APIAccess.GrabNotes(key, post.Item1, post.Item2, beforeTimestamp);
response = await FetchNotesPage(post, beforeTimestamp);
if (response.statusCode == "TooManyRequests")
{
int retry = response.retryInSeconds > 0 ? response.retryInSeconds : 60;
ApiKeyPool.MarkRateLimited(key, retry);
return "TooManyRequests";
if (response.transientFailure)
{
Console.WriteLine($"[Skip] {post.Item1}/{post.Item2} — {response.statusCode} on page {page} after {TransientBackoffSeconds.Length} retries");
return "Transient";
}
if (response.meta?.status != 429)
ApiKeyPool.MarkAvailable(key);
if (IsNotFound(response))
{
@@ -1268,7 +1288,11 @@ if (shouldInsert)
return "UNKNOWN";
}
static async Task CollectNotes(string outPath, bool withoutNotesOnly = true, DateTime? beforeDate = null, bool managedRun = false)
// Consecutive transient failures that mean the API edge is rejecting traffic wholesale rather than
// blipping on one post. Past this, skipping post-by-post would just hammer a closed door.
const int MaxConsecutiveTransient = 10;
static async Task<int> CollectNotes(string outPath, bool withoutNotesOnly = true, DateTime? beforeDate = null, bool managedRun = false)
{
List<Tuple<string, long, long, long>> posts = DataAccess.GetPosts(withoutNotesOnly, beforeDate);
@@ -1277,9 +1301,14 @@ if (shouldInsert)
// post that keeps returning FAILURE/UNKNOWN. Successful/NotFound posts drop out via the DB filter anyway.
HashSet<(string, long)> attempted = new HashSet<(string, long)>();
int skipped = 0;
int consecutiveTransient = 0;
RateLimiter limiter = new SlidingWindowRateLimiter(new SlidingWindowRateLimiterOptions
{
PermitLimit = 300,
// 1/sec average. Sustained higher rates draw CDN-level 403s that the API's own rate-limit
// headers never warn about, so this sits well under the per-key quota on purpose.
PermitLimit = 60,
QueueProcessingOrder = QueueProcessingOrder.OldestFirst,
QueueLimit = 1,
Window = TimeSpan.FromMinutes(1),
@@ -1304,43 +1333,60 @@ if (shouldInsert)
ApiKeyPool.SleepUntilAnyAvailable(30);
string status;
using RateLimitLease lease = limiter.AttemptAcquire(1);
if (lease.IsAcquired)
{
Console.WriteLine("{0,32} - {1,15} - {2}", post.Item1, post.Item2, DateTimeOffset.FromUnixTimeSeconds(post.Item4).ToString());
status = await GrabNotes(post);
}
else
// Wait for a permit rather than giving up on one: the limiter paces the loop, it is
// not a failure condition. Only one acquire is ever pending, so QueueLimit = 1 suffices.
using RateLimitLease lease = await limiter.AcquireAsync(1);
if (!lease.IsAcquired)
{
Console.WriteLine("!@@@@@@@ - Rate Limited Exceeded: No Lease Available");
return; // throttle: abort without completing the run so a later launch resumes
return 3; // abort without completing the run so a later launch resumes
}
if (status == "Success")
Console.WriteLine("{0,32} - {1,15} - {2}", post.Item1, post.Item2, DateTimeOffset.FromUnixTimeSeconds(post.Item4).ToString());
string status = await GrabNotes(post);
if (status == "Transient")
{
// Retries in GrabNotes are already exhausted. Skip the post so the pass can make
// progress; it stays unmarked in the DB, so the next launch picks it up again.
attempted.Add((post.Item1, post.Item2));
DataAccess.UpdatePostMarkNotesCollected(post.Item1, post.Item2);
}
else if (status == "NotFound")
{
attempted.Add((post.Item1, post.Item2));
Console.WriteLine("GrabNotes Result: NotFound");
DataAccess.UpdatePostMarkNotFound(post.Item1, post.Item2);
}
else if (status == "TooManyRequests")
{
// Throttle, not a real per-post failure: don't consume this post's single attempt.
// Abort the pass without completing so a later launch resumes against the same cutoff.
Console.WriteLine("GrabNotes Result: TooManyRequests - pausing run; relaunch to resume.");
return;
skipped++;
consecutiveTransient++;
if (consecutiveTransient >= MaxConsecutiveTransient)
{
Console.WriteLine($"[Abort] {consecutiveTransient} consecutive transient failures - the API edge is rejecting traffic. Pausing run; relaunch to resume. ({skipped} post(s) skipped)");
return 3;
}
}
else
{
// FAILURE / UNKNOWN: count as attempted so the pass can finish instead of retrying forever.
attempted.Add((post.Item1, post.Item2));
Console.WriteLine("GrabNotes Result: " + status);
consecutiveTransient = 0;
if (status == "Success")
{
attempted.Add((post.Item1, post.Item2));
DataAccess.UpdatePostMarkNotesCollected(post.Item1, post.Item2);
}
else if (status == "NotFound")
{
attempted.Add((post.Item1, post.Item2));
Console.WriteLine("GrabNotes Result: NotFound");
DataAccess.UpdatePostMarkNotFound(post.Item1, post.Item2);
}
else if (status == "TooManyRequests")
{
// Throttle, not a real per-post failure: don't consume this post's single attempt.
// Abort the pass without completing so a later launch resumes against the same cutoff.
Console.WriteLine($"GrabNotes Result: TooManyRequests - pausing run; relaunch to resume. ({skipped} post(s) skipped)");
return 3;
}
else
{
// FAILURE / UNKNOWN: count as attempted so the pass can finish instead of retrying forever.
attempted.Add((post.Item1, post.Item2));
Console.WriteLine("GrabNotes Result: " + status);
}
}
// Re-fetch the updated list after processing the current post
@@ -1354,14 +1400,42 @@ if (shouldInsert)
DataAccess.CompleteCollectRun();
Console.WriteLine("Full re-check run complete.");
}
if (skipped > 0)
{
Console.WriteLine($"Pass finished with {skipped} post(s) skipped after transient failures; relaunch to retry them.");
return 3;
}
return 0;
}
catch (Exception ex)
{
Console.WriteLine(ex.ToString());
return 1;
}
}
// Field prefixes TraverseDirectory recognizes as the start of a new record field.
// Used to know where a multi-line Body/Downloaded files value ends.
private static readonly string[] TraverseDirectoryFieldPrefixes = new[]
{
"Post id:", "Reblog url:", "Reblog name:", "Reblog root url:", "Downloaded files:",
"Reblog key:", "Date:", "Body:", "Post url:", "Answer:", "Audio Caption:", "Blog Name:",
"Link:", "Photo Caption:", "Photo url:", "Question:", "Quote:", "Slug:", "Summary:",
"Tags:", "Title:"
};
private static bool IsTraverseDirectoryFieldLine(string line)
{
foreach (var prefix in TraverseDirectoryFieldPrefixes)
{
if (line.StartsWith(prefix, StringComparison.OrdinalIgnoreCase)) return true;
}
return false;
}
static void TraverseDirectory(string path, string outPath, List<string> contains, ref int postsAdded, string blogName = "", string startFromBlogName = "", bool logRecordImports = false)
{
DateTime directoryStart = DateTime.Now;
@@ -1398,8 +1472,10 @@ if (shouldInsert)
var urls = new List<string>();
var reblog = new ReblogRecord();
foreach (string line in File.ReadLines(file))
string[] fileLines = File.ReadAllLines(file);
for (int lineIndex = 0; lineIndex < fileLines.Length; lineIndex++)
{
string line = fileLines[lineIndex];
if (line.StartsWith("Post id:", StringComparison.OrdinalIgnoreCase))
{
if (reblog.reblogName != "." && reblog.postID != "." && reblog.date != "." && reblog.reblogURL != "." && reblog.downloadedFiles == ".")
@@ -1420,7 +1496,7 @@ if (shouldInsert)
DataAccess.AddPost(curDir, long.Parse(reblog.postID), reblog.reblogURL, reblog.date, reblog.postURL, reblog.slug, reblog.reblogKey,
reblog.reblogName, reblog.summary, reblog.quote, reblog.body, reblog.tags, reblog.link, reblog.photoURL,
reblog.photoCaption, reblog.downloadedFiles, reblog.audioCaption, reblog.question, reblog.answer,
reblog.title, false);
reblog.title, false, rootURL: reblog.rootURL);
recordImportStopwatch.Stop();
postsAdded++;
@@ -1445,9 +1521,21 @@ if (shouldInsert)
{
reblog.reblogName = line.Substring(13).Trim();
}
if (line.StartsWith(@"Reblog root url:", StringComparison.OrdinalIgnoreCase))
{
reblog.rootURL = line.Substring(16).Trim();
}
if (line.StartsWith(@"Downloaded files:", StringComparison.OrdinalIgnoreCase))
{
reblog.downloadedFiles = line.Substring(17).Trim();
var valueLines = new List<string> { line.Substring(17).Trim() };
int nextLineIndex = lineIndex + 1;
while (nextLineIndex < fileLines.Length && !IsTraverseDirectoryFieldLine(fileLines[nextLineIndex]))
{
valueLines.Add(fileLines[nextLineIndex]);
nextLineIndex++;
}
reblog.downloadedFiles = string.Join("\n", valueLines).Trim();
lineIndex = nextLineIndex - 1;
}
if (line.StartsWith(@"Reblog key:", StringComparison.OrdinalIgnoreCase))
{
@@ -1459,7 +1547,15 @@ if (shouldInsert)
}
if (line.StartsWith(@"Body:", StringComparison.OrdinalIgnoreCase))
{
reblog.body = line.Substring(6).Trim();
var valueLines = new List<string> { line.Substring(6).Trim() };
int nextLineIndex = lineIndex + 1;
while (nextLineIndex < fileLines.Length && !IsTraverseDirectoryFieldLine(fileLines[nextLineIndex]))
{
valueLines.Add(fileLines[nextLineIndex]);
nextLineIndex++;
}
reblog.body = string.Join("\n", valueLines).Trim();
lineIndex = nextLineIndex - 1;
}
if (line.StartsWith(@"Post url:", StringComparison.OrdinalIgnoreCase))
{
@@ -1548,7 +1644,7 @@ if (shouldInsert)
DataAccess.AddPost(curDir, long.Parse(reblog.postID), reblog.reblogURL, reblog.date, reblog.postURL, reblog.slug, reblog.reblogKey,
reblog.reblogName, reblog.summary, reblog.quote, reblog.body, reblog.tags, reblog.link, reblog.photoURL,
reblog.photoCaption, reblog.downloadedFiles, reblog.audioCaption, reblog.question, reblog.answer,
reblog.title, true);
reblog.title, true, rootURL: reblog.rootURL);
recordImportStopwatch.Stop();
postsAdded++;
+4
View File
@@ -85,6 +85,10 @@ namespace URLNotesGrabberCORE
public int retryInSeconds { get; set; }
public string rawJson { get; set; }
// The request never reached the Tumblr API (transport error, or an edge/CDN response with a
// non-JSON body). Says nothing about the post, so the caller should retry rather than fail it.
public bool transientFailure { get; set; }
}
// Classes for Posts API endpoint (for reply_text)