Skip to content

Fix infostream_to_segments.py to parse compact segment descriptors #597

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
mikemccand merged 1 commit into from
Jul 23, 2026
Merged

Fix infostream_to_segments.py to parse compact segment descriptors #597

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
15 changes: 14 additions & 1 deletion src/python/infostream_to_segments.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -24,6 +24,11 @@
import re
import sys

# Fallback when parsing a log that doesn't contain the verbose segment details
# (e.g. truncated log missing the init/flush/merge birth line). We assume
# "merge" since long-lived indices are dominated by merged segments.
DEFAULT_SEGMENT_SOURCE = ("merge", [])


# thank you Claude:
def parse_timestamp(timestamp_str: str) -> datetime.datetime | None:
Expand Down Expand Up @@ -125,7 +130,13 @@ def parse_timestamp(timestamp_str: str) -> datetime.datetime | None:

# e.g. typically many on one line like this: _f(11.0.0):C49781:[diagnostics={os.arch=amd64, os.version=6.12.4-arch1-1, lucene.version=11.0.0, source=flush, timestamp=1738326951546, java.runtime.version=23, java.vendor=Arch Linux, os=Linux}]:[attributes={Lucene90StoredFieldsFormat.mode=BEST_SPEED}] :id=o6ynrzazp1d8kt6wycnpppvw
# or (with static index sort): IFD 211 [2025-02-06T18:26:20.518261400Z; GCR-Writer-1-thread-6]: now checkpoint "_a(9.11.2):C14171/34:[indexSort=<int: "marketplaceid"> missingValue=-1,<long: "asin_ve_sort_value">! missingValue=0,<string: "ve-family-id"> missingValue=SortField.STRING_FIRST,<string: "docid"> missingValue=SortField.STRING_FIRST]:[diagnostics={timestamp=1738866380509, java.runtime.version=23.0.2+7, java.vendor=Amazon.com Inc., os=Linux, os.arch=aarch64, os.version=4.14.355-275.582.amzn2.aarch64, lucene.version=9.11.2, source=flush}]:[attributes={Lucene90StoredFieldsFormat.mode=BEST_SPEED}]:delGen=1 :id=dz4cethjje165l9ysayg0m9f1" [1 segments ; isCommit = false]
re_segment_desc = re.compile(r"_(.*?)\(.*?\):([cC])(\d+)(/\d+)?:\[(indexSort=.*?)?(diagnostics=\{.*?\})\]:\[attributes=(\{.*?\})\](:delGen=\d+)? :id=[0-9a-z]+?")
# Since https://github.com/apache/lucene/pull/15885, IndexWriter emits verbose
# segment details (diagnostics, attributes, id) only once — at init, flush, or
# merge time. Subsequent references use a compact form, e.g.
# `_4(10.4.0):C8782/109`, because the verbose version was already logged once
# for this segment. The trailer below is wrapped in an optional non-capturing
# group so both forms match; group numbering (1-8) is preserved.
re_segment_desc = re.compile(r"_(.*?)\(.*?\):([cC])(\d+)(/\d+)?(?::\[(indexSort=.*?)?(diagnostics=\{.*?\})\]:\[attributes=(\{.*?\})\](:delGen=\d+)? :id=[0-9a-z]+)?")
Comment thread
jerinthomas1404 marked this conversation as resolved.


def sec_to_time_delta(td_sec):
Expand Down Expand Up @@ -410,6 +421,8 @@ def parse_seg_details(s):
source = ("merge", [])
elif "source=flush" in diagnostics:
source = "flush"
elif diagnostics == "":
source = DEFAULT_SEGMENT_SOURCE
else:
raise RuntimeError(f"seg {segment_name}: a new source in {diagnostics}?")

Expand Down
Loading