Add RC1 regression benchmark artifacts
Add benchmark artifacts created during the first RC1 robustness evaluation of the Meeting Lab pipeline. Included benchmark sets: - hardware_experimental_qwen35_9b - meeting_context_v1 - progeo_meeting_20260804_083849 - progeo_meeting_context_v1_20260804_110913 - progeo_meeting_rc1_20260804_121502 These artifacts document the evolution of the pipeline during the implementation and verification of BUG-009, BUG-010 and BUG-011. The benchmark data provide reproducible real-life regression cases for future development and allow quality comparisons across pipeline revisions. Current benchmark policy: During the active development phase, representative benchmark artifacts are intentionally versioned to preserve reproducibility and simplify regression analysis. Benchmark artifacts are considered part of the engineering evidence rather than temporary build output. Benchmark retention strategy will be revisited once the Meeting Assistant reaches production maturity.
This commit is contained in:
+264
@@ -0,0 +1,264 @@
|
||||
{
|
||||
"groups": [
|
||||
{
|
||||
"canonical_text": "Wir haben bei der 100%-Variante insgesamt 126.000 Laufmeter extrudiert.",
|
||||
"source_item_ids": ["fact_0001", "fact_0015"],
|
||||
"merge_reason": "Both items state the exact same production volume (126,000 meters) for the 100% RC variant. fact_0001 uses 'extrudiert' and fact_0015 uses 'hatten', but they refer to the identical factual proposition regarding output quantity."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Die Stäbe waren leider etwas breiter und auch zu dünn.",
|
||||
"source_item_ids": ["fact_0002", "fact_0012"],
|
||||
"merge_reason": "Both items describe the physical dimensions of the bars (broader and thinner). fact_0012 adds context about them not breaking, but shares the core factual proposition regarding their width and thickness."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Das RC-Material hatte ungefähr eine dreifach niedrigere Viskosität.",
|
||||
"source_item_ids": ["fact_0003"],
|
||||
"merge_reason": "Singleton. Describes a specific property of the material without direct equivalent in other items."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Die Anlagen können technisch gesehen nur eine Hauptkomponente und eigentlich eine Zusatzkomponente verarbeiten.",
|
||||
"source_item_ids": ["fact_0004"],
|
||||
"merge_reason": "Singleton. Describes the technical limitation of the machinery."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Die dünneren Stäbe resultieren daraus, dass es kein Virgin-Material war.",
|
||||
"source_item_ids": ["fact_0005"],
|
||||
"merge_reason": "Singleton. Establishes a causal link between material type and bar thickness."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Im Vergleich zum 4040er-Stab waren wir bei der Zugfestigkeit um 23% unter dem üblichen Wert.",
|
||||
"source_item_ids": ["fact_0006"],
|
||||
"merge_reason": "Singleton. Specific tensile strength deviation for a specific bar type."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Im Vergleich zum 30-30er-Stab war die Zugfestigkeit um 8% geringer.",
|
||||
"source_item_ids": ["fact_0007"],
|
||||
"merge_reason": "Singleton. Specific tensile strength deviation for a different bar type."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Die Dehnung lag im Vergleich zum 30-30er-Stab um 22% niedriger.",
|
||||
"source_item_ids": ["fact_0008"],
|
||||
"merge_reason": "Singleton. Specific elongation deviation for the same bar type as fact_0007."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Das RC-Material hat im Grunde die gleiche Genese, und bei einem einmaligen Durchlauf sehen wir keinen Unterschied.",
|
||||
"source_item_ids": ["fact_0009"],
|
||||
"merge_reason": "Singleton. Describes the origin and immediate result of a single pass."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Störstellen beim Zugversuch verringern den wirksamen Querschnitt und führen zum Reißen.",
|
||||
"source_item_ids": ["fact_0010"],
|
||||
"merge_reason": "Singleton. Explains the mechanical failure mechanism."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Wir haben die Ergebnisse nur gegen unsere Datenblätter verglichen.",
|
||||
"source_item_ids": ["fact_0011"],
|
||||
"merge_reason": "Singleton. Describes the basis of comparison used in testing."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Die Stäbe sind nicht mehr gerissen, außer dass sie leider wieder etwas breiter und dünner waren.",
|
||||
"source_item_ids": ["fact_0012"],
|
||||
"merge_reason": "Note: fact_0012 was merged with fact_0002. This entry is a duplicate in the source list or refers to the same event described there."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Bei 20% RC-Anteil waren die Festigkeiten vergleichbar mit dem 40-40er Gitter, bei 100% RC-Anteil lagen sie sogar unter dem 30-30er Gitter.",
|
||||
"source_item_ids": ["fact_0013"],
|
||||
"merge_reason": "Singleton. Compares strength across different material percentages and grid types."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Bei 20% Regenumann waren wir etwas besser bei der Brustspannung.",
|
||||
"source_item_ids": ["fact_0014"],
|
||||
"merge_reason": "Singleton. Specific performance note for a specific material mix (likely typo in source 'Regenumann' vs context, treated as distinct fact)."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Bei 20% RC-Anteil hatten wir 80.000 Laufmeter.",
|
||||
"source_item_ids": ["fact_0016"],
|
||||
"merge_reason": "Singleton. Specific production volume for the 20% variant."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Wir haben ungefähr sechs Tonnen Material reingegeben und eine Tonne unten rausbekommen.",
|
||||
"source_item_ids": ["fact_0017"],
|
||||
"merge_reason": "Singleton. Describes the input/output mass balance of a process step."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Wir haben bewiesen, dass es im hundertprozentigen Umlauf geht.",
|
||||
"source_item_ids": ["fact_0018"],
|
||||
"merge_reason": "Singleton. States the successful proof of concept for 100% recycling."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Die realistischere Variante ist eine prozentuale Zumischung zu Virgin Material.",
|
||||
"source_item_ids": ["fact_0019"],
|
||||
"merge_reason": "Singleton. States a strategic conclusion or preference regarding material mixing."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Wir haben ungefähr 30 Kilo an Telalux geschickt (genauer Umfang unklar).",
|
||||
"source_item_ids": ["fact_0020"],
|
||||
"merge_reason": "Singleton. Describes the quantity of material sent to a specific supplier."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Der Lieferant hat zwei verschiedene Dosierungen an seinem Marker rein dosiert.",
|
||||
"source_item_ids": ["fact_0021"],
|
||||
"merge_reason": "Singleton. Describes actions taken by the external party (supplier)."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Wir haben ungefähr 30 Kilo Material, davon ist der genaue verbleibende Umfang unklar.",
|
||||
"source_item_ids": ["fact_0022"],
|
||||
"merge_reason": "Singleton. Reiteration of the quantity held by 'us', distinct from what was sent."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Ich neige dazu, die Ballen bei Cockman jetzt zu entsorgen.",
|
||||
"source_item_ids": ["fact_0023"],
|
||||
"merge_reason": "Singleton. Describes an intention regarding waste management of specific bales."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Das Material ist für ein Geogitter recht alt, etwa 20 Jahre.",
|
||||
"source_item_ids": ["fact_0024"],
|
||||
"merge_reason": "Singleton. Describes the age/condition of a specific material batch."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Polyester wird etwas anders behandelt als Polypropylen.",
|
||||
"source_item_ids": ["fact_0025"],
|
||||
"merge_reason": "Singleton. General statement on processing differences between materials."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Wir haben Muster geschnitten, die nicht so aussahen wie in der Mulde und dreckigeres Zeug dabei war.",
|
||||
"source_item_ids": ["fact_0026"],
|
||||
"merge_reason": "Singleton. Describes the condition of cut samples versus remaining material."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Ein typischer Alterungsversuch prüft eine Nullprobe, setzt sie Umweltbedingungen aus (beschleunigt), und prüft danach wieder Dehnung und Festigkeit.",
|
||||
"source_item_ids": ["fact_0027"],
|
||||
"merge_reason": "Singleton. Describes the standard procedure for aging tests."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Man findet erfahrungsgemäß kaum viele optisch unversehrte Stücke, sondern eher Stablose oder Knotenlose.",
|
||||
"source_item_ids": ["fact_0028"],
|
||||
"merge_reason": "Singleton. Describes the typical quality distribution of samples."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Wir kommen in unserem Produktionsprozess mit einer Tonne schon ein ganzes Stück weit.",
|
||||
"source_item_ids": ["fact_0029", "fact_0032"],
|
||||
"merge_reason": "Both items state that processing one tonne represents significant progress. fact_0032 mentions 'four tonnes' in a different context (estimation of stock), but the core proposition about achieving results with one tonne is shared."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Die Ratio ist bei dem Polyester-Gitter besser als beim PP-Kombi-Grid, weil viel Gewicht im Fließstoff steckte.",
|
||||
"source_item_ids": ["fact_0030"],
|
||||
"merge_reason": "Singleton. Compares product ratios based on material composition."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Wir haben für die ersten zwei Mulden allein an Transportkosten fast 5000 Euro bezahlt.",
|
||||
"source_item_ids": ["fact_0031"],
|
||||
"merge_reason": "Singleton. Describes specific costs incurred for transport."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Wir haben versucht, rückwärts eine Rückwärtssuche zu machen.",
|
||||
"source_item_ids": ["fact_0033"],
|
||||
"merge_reason": "Singleton. Describes a methodological approach taken in investigation."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Wir haben das so gefahren, wie unser Standardprodukt auch fahren.",
|
||||
"source_item_ids": ["fact_0034"],
|
||||
"merge_reason": "Singleton. States that the process followed standard procedures."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Bei Kockmann nutzen wir ein 20er Trommelsieb, da es dort vorhanden war.",
|
||||
"source_item_ids": ["fact_0035"],
|
||||
"merge_reason": "Singleton. Describes equipment usage at a specific location."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Wir haben versucht, den Siebschnitt noch mal zu optimieren.",
|
||||
"source_item_ids": ["fact_0036"],
|
||||
"merge_reason": "Singleton. States an optimization effort regarding sieving."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Wir haben für diesen Prozess ziemlich viel Energieaufwand reingeschickt.",
|
||||
"source_item_ids": ["fact_0037"],
|
||||
"merge_reason": "Singleton. Describes the energy consumption of a process step."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Unser Aschegehalt war sogar größer als beim ersten Großversuch.",
|
||||
"source_item_ids": ["fact_0038"],
|
||||
"merge_reason": "Singleton. Compares ash content between current and previous large-scale trials."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Es gibt teilweise bisher noch nicht so erfolgreiche Rückmeldungen, die erst im August nach der Urlaubszeit kommen.",
|
||||
"source_item_ids": ["fact_0039"],
|
||||
"merge_reason": "Singleton. Describes the status and timing of pending feedback."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Wir können neues Material mit einem Anteil von 20 bis zu 100 Prozent generieren.",
|
||||
"source_item_ids": ["fact_0040"],
|
||||
"merge_reason": "Singleton. States the capability range for material generation."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Ich habe gestern ein Angebot für diese Versuche bekommen, das 30.000 Euro kostet.",
|
||||
"source_item_ids": ["fact_0041"],
|
||||
"merge_reason": "Singleton. Describes a specific cost offer received."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Wir haben dankend abgewogen und entschieden, dass wir das Angebot vielleicht doch nicht brauchen.",
|
||||
"source_item_ids": ["fact_0042", "fact_0043"],
|
||||
"merge_reason": "Both items describe the decision-making process regarding the offer. fact_0042 states they weighed it and decided not to need it; fact_0043 confirms this hesitation/decision."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Das Polymer ist nachher nicht mehr unterscheidbar von einem aus Erdöl gewonnenen Polymer, da dem Polymer egal ist, wo das Monomer hergekommen ist.",
|
||||
"source_item_ids": ["fact_0044"],
|
||||
"merge_reason": "Singleton. Explains the chemical indistinguishability of recycled polymers."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Das Material wird tatsächlich noch mal weiter genutzt worden sein.",
|
||||
"source_item_ids": ["fact_0045"],
|
||||
"merge_reason": "Singleton. States that material will be reused."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Wir haben bewiesen, was zu beweisen war auf unserer Seite.",
|
||||
"source_item_ids": ["fact_0046"],
|
||||
"merge_reason": "Singleton. Declares the completion of proof obligations."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Ich gehe in Urlaub vom 2. bis zum 11.",
|
||||
"source_item_ids": ["fact_0047", "fact_0051"],
|
||||
"merge_reason": "Both items refer to the same vacation period (August/September context). fact_0047 gives dates, fact_0051 confirms presence on 30th. Merged as they describe the same event timeline."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Ich fliege am 31., sitze ich im Flugzeug.",
|
||||
"source_item_ids": ["fact_0048", "fact_0049"],
|
||||
"merge_reason": "Both items describe the departure flight on the 31st. fact_0049 is a fragment 'Ich bin ab' which semantically equates to departing."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Wir stellen die Ergebnisse am ersten Tag direkt mit einem Vortrag vor.",
|
||||
"source_item_ids": ["fact_0050"],
|
||||
"merge_reason": "Singleton. Describes the presentation schedule for results."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Der CET-Bebeitrag steht Mitte August an und wird wahrscheinlich im Konsortium abgestimmt.",
|
||||
"source_item_ids": ["fact_0052"],
|
||||
"merge_reason": "Singleton. Describes the deadline and process for a specific contribution."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Alle Verbundprojekte werden vorgestellt, mit einem Umfang von 18.000 Zeichen pro Artikel.",
|
||||
"source_item_ids": ["fact_0053"],
|
||||
"merge_reason": "Singleton. Describes the constraints for project abstracts."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Der Beitrag ist bis Ende oder Mitte September abzugeben.",
|
||||
"source_item_ids": ["fact_0054"],
|
||||
"merge_reason": "Singleton. States the deadline for submission of a contribution."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Das wird bei diesem Extruder schlechter gehen als am großen, weil dieser aufgrund seiner Größe nicht so gut mischt.",
|
||||
"source_item_ids": ["fact_0055"],
|
||||
"merge_reason": "Singleton. Compares mixing performance between extruders."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Der Extruder ist beim Extudieren kaputt gegangen, da haben wir nichts mehr zu schaffen.",
|
||||
"source_item_ids": ["fact_0056"],
|
||||
"merge_reason": "Singleton. Describes equipment failure and cessation of production."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Die E-Mail war vom 17.07.",
|
||||
"source_item_ids": ["fact_0057"],
|
||||
"merge_reason": "Singleton. States the date of a specific communication."
|
||||
}
|
||||
]
|
||||
}
|
||||
+43
@@ -0,0 +1,43 @@
|
||||
{
|
||||
"repair_engine_available_as_importable_package": false,
|
||||
"repair_basis": "docs/constraint-repair.md generic validator interface and V1 deterministic operations",
|
||||
"llm_used": false,
|
||||
"retry_count": 0,
|
||||
"runtime_seconds": 0.002917,
|
||||
"input_model_output": "samples/benchmarks/progeo_meeting_20260804_083849/semantic_consolidator/raw_model_response.txt",
|
||||
"validator_before": "samples/benchmarks/progeo_meeting_20260804_083849/semantic_consolidator/validator_before.json",
|
||||
"repaired_model_groups": "samples/benchmarks/progeo_meeting_20260804_083849/semantic_consolidator/repaired_model_groups.json",
|
||||
"repaired_consolidated_output": "samples/benchmarks/progeo_meeting_20260804_083849/semantic_consolidator/repaired_consolidated_extractions.json",
|
||||
"validator_after": "samples/benchmarks/progeo_meeting_20260804_083849/semantic_consolidator/validator_after.json",
|
||||
"original_validator_error": "Source item IDs appear in multiple groups: ['fact_0012']",
|
||||
"structural_changes": [
|
||||
{
|
||||
"operation": "remove_duplicate_identifier_occurrence",
|
||||
"id": "fact_0012",
|
||||
"collection_pointer": "/groups",
|
||||
"id_list_key": "source_item_ids",
|
||||
"removed_occurrence": {
|
||||
"item_index": 11,
|
||||
"id_index": 0
|
||||
},
|
||||
"kept_occurrence": {
|
||||
"item_index": 1,
|
||||
"id_index": 1
|
||||
}
|
||||
},
|
||||
{
|
||||
"operation": "remove_empty_group",
|
||||
"collection_pointer": "/groups",
|
||||
"item_index": 11,
|
||||
"removed_group_canonical_text": "Die Stäbe sind nicht mehr gerissen, außer dass sie leider wieder etwas breiter und dünner waren.",
|
||||
"removed_group_merge_reason": "Note: fact_0012 was merged with fact_0002. This entry is a duplicate in the source list or refers to the same event described there."
|
||||
}
|
||||
],
|
||||
"fact_0012_occurrences_after": [
|
||||
{
|
||||
"item_index": 1,
|
||||
"id_index": 1
|
||||
}
|
||||
],
|
||||
"non_fact_categories_unchanged": true
|
||||
}
|
||||
+2656
File diff suppressed because it is too large
Load Diff
+367
@@ -0,0 +1,367 @@
|
||||
{
|
||||
"groups": [
|
||||
{
|
||||
"canonical_text": "Wir haben bei der 100%-Variante insgesamt 126.000 Laufmeter extrudiert.",
|
||||
"source_item_ids": [
|
||||
"fact_0001",
|
||||
"fact_0015"
|
||||
],
|
||||
"merge_reason": "Both items state the exact same production volume (126,000 meters) for the 100% RC variant. fact_0001 uses 'extrudiert' and fact_0015 uses 'hatten', but they refer to the identical factual proposition regarding output quantity."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Die Stäbe waren leider etwas breiter und auch zu dünn.",
|
||||
"source_item_ids": [
|
||||
"fact_0002",
|
||||
"fact_0012"
|
||||
],
|
||||
"merge_reason": "Both items describe the physical dimensions of the bars (broader and thinner). fact_0012 adds context about them not breaking, but shares the core factual proposition regarding their width and thickness."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Das RC-Material hatte ungefähr eine dreifach niedrigere Viskosität.",
|
||||
"source_item_ids": [
|
||||
"fact_0003"
|
||||
],
|
||||
"merge_reason": "Singleton. Describes a specific property of the material without direct equivalent in other items."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Die Anlagen können technisch gesehen nur eine Hauptkomponente und eigentlich eine Zusatzkomponente verarbeiten.",
|
||||
"source_item_ids": [
|
||||
"fact_0004"
|
||||
],
|
||||
"merge_reason": "Singleton. Describes the technical limitation of the machinery."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Die dünneren Stäbe resultieren daraus, dass es kein Virgin-Material war.",
|
||||
"source_item_ids": [
|
||||
"fact_0005"
|
||||
],
|
||||
"merge_reason": "Singleton. Establishes a causal link between material type and bar thickness."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Im Vergleich zum 4040er-Stab waren wir bei der Zugfestigkeit um 23% unter dem üblichen Wert.",
|
||||
"source_item_ids": [
|
||||
"fact_0006"
|
||||
],
|
||||
"merge_reason": "Singleton. Specific tensile strength deviation for a specific bar type."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Im Vergleich zum 30-30er-Stab war die Zugfestigkeit um 8% geringer.",
|
||||
"source_item_ids": [
|
||||
"fact_0007"
|
||||
],
|
||||
"merge_reason": "Singleton. Specific tensile strength deviation for a different bar type."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Die Dehnung lag im Vergleich zum 30-30er-Stab um 22% niedriger.",
|
||||
"source_item_ids": [
|
||||
"fact_0008"
|
||||
],
|
||||
"merge_reason": "Singleton. Specific elongation deviation for the same bar type as fact_0007."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Das RC-Material hat im Grunde die gleiche Genese, und bei einem einmaligen Durchlauf sehen wir keinen Unterschied.",
|
||||
"source_item_ids": [
|
||||
"fact_0009"
|
||||
],
|
||||
"merge_reason": "Singleton. Describes the origin and immediate result of a single pass."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Störstellen beim Zugversuch verringern den wirksamen Querschnitt und führen zum Reißen.",
|
||||
"source_item_ids": [
|
||||
"fact_0010"
|
||||
],
|
||||
"merge_reason": "Singleton. Explains the mechanical failure mechanism."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Wir haben die Ergebnisse nur gegen unsere Datenblätter verglichen.",
|
||||
"source_item_ids": [
|
||||
"fact_0011"
|
||||
],
|
||||
"merge_reason": "Singleton. Describes the basis of comparison used in testing."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Bei 20% RC-Anteil waren die Festigkeiten vergleichbar mit dem 40-40er Gitter, bei 100% RC-Anteil lagen sie sogar unter dem 30-30er Gitter.",
|
||||
"source_item_ids": [
|
||||
"fact_0013"
|
||||
],
|
||||
"merge_reason": "Singleton. Compares strength across different material percentages and grid types."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Bei 20% Regenumann waren wir etwas besser bei der Brustspannung.",
|
||||
"source_item_ids": [
|
||||
"fact_0014"
|
||||
],
|
||||
"merge_reason": "Singleton. Specific performance note for a specific material mix (likely typo in source 'Regenumann' vs context, treated as distinct fact)."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Bei 20% RC-Anteil hatten wir 80.000 Laufmeter.",
|
||||
"source_item_ids": [
|
||||
"fact_0016"
|
||||
],
|
||||
"merge_reason": "Singleton. Specific production volume for the 20% variant."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Wir haben ungefähr sechs Tonnen Material reingegeben und eine Tonne unten rausbekommen.",
|
||||
"source_item_ids": [
|
||||
"fact_0017"
|
||||
],
|
||||
"merge_reason": "Singleton. Describes the input/output mass balance of a process step."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Wir haben bewiesen, dass es im hundertprozentigen Umlauf geht.",
|
||||
"source_item_ids": [
|
||||
"fact_0018"
|
||||
],
|
||||
"merge_reason": "Singleton. States the successful proof of concept for 100% recycling."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Die realistischere Variante ist eine prozentuale Zumischung zu Virgin Material.",
|
||||
"source_item_ids": [
|
||||
"fact_0019"
|
||||
],
|
||||
"merge_reason": "Singleton. States a strategic conclusion or preference regarding material mixing."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Wir haben ungefähr 30 Kilo an Telalux geschickt (genauer Umfang unklar).",
|
||||
"source_item_ids": [
|
||||
"fact_0020"
|
||||
],
|
||||
"merge_reason": "Singleton. Describes the quantity of material sent to a specific supplier."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Der Lieferant hat zwei verschiedene Dosierungen an seinem Marker rein dosiert.",
|
||||
"source_item_ids": [
|
||||
"fact_0021"
|
||||
],
|
||||
"merge_reason": "Singleton. Describes actions taken by the external party (supplier)."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Wir haben ungefähr 30 Kilo Material, davon ist der genaue verbleibende Umfang unklar.",
|
||||
"source_item_ids": [
|
||||
"fact_0022"
|
||||
],
|
||||
"merge_reason": "Singleton. Reiteration of the quantity held by 'us', distinct from what was sent."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Ich neige dazu, die Ballen bei Cockman jetzt zu entsorgen.",
|
||||
"source_item_ids": [
|
||||
"fact_0023"
|
||||
],
|
||||
"merge_reason": "Singleton. Describes an intention regarding waste management of specific bales."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Das Material ist für ein Geogitter recht alt, etwa 20 Jahre.",
|
||||
"source_item_ids": [
|
||||
"fact_0024"
|
||||
],
|
||||
"merge_reason": "Singleton. Describes the age/condition of a specific material batch."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Polyester wird etwas anders behandelt als Polypropylen.",
|
||||
"source_item_ids": [
|
||||
"fact_0025"
|
||||
],
|
||||
"merge_reason": "Singleton. General statement on processing differences between materials."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Wir haben Muster geschnitten, die nicht so aussahen wie in der Mulde und dreckigeres Zeug dabei war.",
|
||||
"source_item_ids": [
|
||||
"fact_0026"
|
||||
],
|
||||
"merge_reason": "Singleton. Describes the condition of cut samples versus remaining material."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Ein typischer Alterungsversuch prüft eine Nullprobe, setzt sie Umweltbedingungen aus (beschleunigt), und prüft danach wieder Dehnung und Festigkeit.",
|
||||
"source_item_ids": [
|
||||
"fact_0027"
|
||||
],
|
||||
"merge_reason": "Singleton. Describes the standard procedure for aging tests."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Man findet erfahrungsgemäß kaum viele optisch unversehrte Stücke, sondern eher Stablose oder Knotenlose.",
|
||||
"source_item_ids": [
|
||||
"fact_0028"
|
||||
],
|
||||
"merge_reason": "Singleton. Describes the typical quality distribution of samples."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Wir kommen in unserem Produktionsprozess mit einer Tonne schon ein ganzes Stück weit.",
|
||||
"source_item_ids": [
|
||||
"fact_0029",
|
||||
"fact_0032"
|
||||
],
|
||||
"merge_reason": "Both items state that processing one tonne represents significant progress. fact_0032 mentions 'four tonnes' in a different context (estimation of stock), but the core proposition about achieving results with one tonne is shared."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Die Ratio ist bei dem Polyester-Gitter besser als beim PP-Kombi-Grid, weil viel Gewicht im Fließstoff steckte.",
|
||||
"source_item_ids": [
|
||||
"fact_0030"
|
||||
],
|
||||
"merge_reason": "Singleton. Compares product ratios based on material composition."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Wir haben für die ersten zwei Mulden allein an Transportkosten fast 5000 Euro bezahlt.",
|
||||
"source_item_ids": [
|
||||
"fact_0031"
|
||||
],
|
||||
"merge_reason": "Singleton. Describes specific costs incurred for transport."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Wir haben versucht, rückwärts eine Rückwärtssuche zu machen.",
|
||||
"source_item_ids": [
|
||||
"fact_0033"
|
||||
],
|
||||
"merge_reason": "Singleton. Describes a methodological approach taken in investigation."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Wir haben das so gefahren, wie unser Standardprodukt auch fahren.",
|
||||
"source_item_ids": [
|
||||
"fact_0034"
|
||||
],
|
||||
"merge_reason": "Singleton. States that the process followed standard procedures."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Bei Kockmann nutzen wir ein 20er Trommelsieb, da es dort vorhanden war.",
|
||||
"source_item_ids": [
|
||||
"fact_0035"
|
||||
],
|
||||
"merge_reason": "Singleton. Describes equipment usage at a specific location."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Wir haben versucht, den Siebschnitt noch mal zu optimieren.",
|
||||
"source_item_ids": [
|
||||
"fact_0036"
|
||||
],
|
||||
"merge_reason": "Singleton. States an optimization effort regarding sieving."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Wir haben für diesen Prozess ziemlich viel Energieaufwand reingeschickt.",
|
||||
"source_item_ids": [
|
||||
"fact_0037"
|
||||
],
|
||||
"merge_reason": "Singleton. Describes the energy consumption of a process step."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Unser Aschegehalt war sogar größer als beim ersten Großversuch.",
|
||||
"source_item_ids": [
|
||||
"fact_0038"
|
||||
],
|
||||
"merge_reason": "Singleton. Compares ash content between current and previous large-scale trials."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Es gibt teilweise bisher noch nicht so erfolgreiche Rückmeldungen, die erst im August nach der Urlaubszeit kommen.",
|
||||
"source_item_ids": [
|
||||
"fact_0039"
|
||||
],
|
||||
"merge_reason": "Singleton. Describes the status and timing of pending feedback."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Wir können neues Material mit einem Anteil von 20 bis zu 100 Prozent generieren.",
|
||||
"source_item_ids": [
|
||||
"fact_0040"
|
||||
],
|
||||
"merge_reason": "Singleton. States the capability range for material generation."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Ich habe gestern ein Angebot für diese Versuche bekommen, das 30.000 Euro kostet.",
|
||||
"source_item_ids": [
|
||||
"fact_0041"
|
||||
],
|
||||
"merge_reason": "Singleton. Describes a specific cost offer received."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Wir haben dankend abgewogen und entschieden, dass wir das Angebot vielleicht doch nicht brauchen.",
|
||||
"source_item_ids": [
|
||||
"fact_0042",
|
||||
"fact_0043"
|
||||
],
|
||||
"merge_reason": "Both items describe the decision-making process regarding the offer. fact_0042 states they weighed it and decided not to need it; fact_0043 confirms this hesitation/decision."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Das Polymer ist nachher nicht mehr unterscheidbar von einem aus Erdöl gewonnenen Polymer, da dem Polymer egal ist, wo das Monomer hergekommen ist.",
|
||||
"source_item_ids": [
|
||||
"fact_0044"
|
||||
],
|
||||
"merge_reason": "Singleton. Explains the chemical indistinguishability of recycled polymers."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Das Material wird tatsächlich noch mal weiter genutzt worden sein.",
|
||||
"source_item_ids": [
|
||||
"fact_0045"
|
||||
],
|
||||
"merge_reason": "Singleton. States that material will be reused."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Wir haben bewiesen, was zu beweisen war auf unserer Seite.",
|
||||
"source_item_ids": [
|
||||
"fact_0046"
|
||||
],
|
||||
"merge_reason": "Singleton. Declares the completion of proof obligations."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Ich gehe in Urlaub vom 2. bis zum 11.",
|
||||
"source_item_ids": [
|
||||
"fact_0047",
|
||||
"fact_0051"
|
||||
],
|
||||
"merge_reason": "Both items refer to the same vacation period (August/September context). fact_0047 gives dates, fact_0051 confirms presence on 30th. Merged as they describe the same event timeline."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Ich fliege am 31., sitze ich im Flugzeug.",
|
||||
"source_item_ids": [
|
||||
"fact_0048",
|
||||
"fact_0049"
|
||||
],
|
||||
"merge_reason": "Both items describe the departure flight on the 31st. fact_0049 is a fragment 'Ich bin ab' which semantically equates to departing."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Wir stellen die Ergebnisse am ersten Tag direkt mit einem Vortrag vor.",
|
||||
"source_item_ids": [
|
||||
"fact_0050"
|
||||
],
|
||||
"merge_reason": "Singleton. Describes the presentation schedule for results."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Der CET-Bebeitrag steht Mitte August an und wird wahrscheinlich im Konsortium abgestimmt.",
|
||||
"source_item_ids": [
|
||||
"fact_0052"
|
||||
],
|
||||
"merge_reason": "Singleton. Describes the deadline and process for a specific contribution."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Alle Verbundprojekte werden vorgestellt, mit einem Umfang von 18.000 Zeichen pro Artikel.",
|
||||
"source_item_ids": [
|
||||
"fact_0053"
|
||||
],
|
||||
"merge_reason": "Singleton. Describes the constraints for project abstracts."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Der Beitrag ist bis Ende oder Mitte September abzugeben.",
|
||||
"source_item_ids": [
|
||||
"fact_0054"
|
||||
],
|
||||
"merge_reason": "Singleton. States the deadline for submission of a contribution."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Das wird bei diesem Extruder schlechter gehen als am großen, weil dieser aufgrund seiner Größe nicht so gut mischt.",
|
||||
"source_item_ids": [
|
||||
"fact_0055"
|
||||
],
|
||||
"merge_reason": "Singleton. Compares mixing performance between extruders."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Der Extruder ist beim Extudieren kaputt gegangen, da haben wir nichts mehr zu schaffen.",
|
||||
"source_item_ids": [
|
||||
"fact_0056"
|
||||
],
|
||||
"merge_reason": "Singleton. Describes equipment failure and cessation of production."
|
||||
},
|
||||
{
|
||||
"canonical_text": "Die E-Mail war vom 17.07.",
|
||||
"source_item_ids": [
|
||||
"fact_0057"
|
||||
],
|
||||
"merge_reason": "Singleton. States the date of a specific communication."
|
||||
}
|
||||
]
|
||||
}
|
||||
BIN
Binary file not shown.
+12
@@ -0,0 +1,12 @@
|
||||
{
|
||||
"valid": true,
|
||||
"interface": "generic_constraint_validator_v1",
|
||||
"stage": "semantic_consolidator_model_groups",
|
||||
"source_document": "samples/benchmarks/progeo_meeting_20260804_083849/semantic_consolidator/raw_model_response.txt",
|
||||
"canonicalizer_input": "samples/benchmarks/progeo_meeting_20260804_083849/canonicalizer/canonicalized_extractions.json",
|
||||
"expected_source_id_count": 57,
|
||||
"observed_source_id_occurrence_count": 57,
|
||||
"unique_observed_source_id_count": 57,
|
||||
"non_fact_categories_unchanged": true,
|
||||
"violations": []
|
||||
}
|
||||
+32
@@ -0,0 +1,32 @@
|
||||
{
|
||||
"valid": false,
|
||||
"interface": "generic_constraint_validator_v1",
|
||||
"stage": "semantic_consolidator_model_groups",
|
||||
"source_document": "samples/benchmarks/progeo_meeting_20260804_083849/semantic_consolidator/raw_model_response.txt",
|
||||
"canonicalizer_input": "samples/benchmarks/progeo_meeting_20260804_083849/canonicalizer/canonicalized_extractions.json",
|
||||
"expected_source_id_count": 57,
|
||||
"observed_source_id_occurrence_count": 58,
|
||||
"unique_observed_source_id_count": 57,
|
||||
"non_fact_categories_unchanged": null,
|
||||
"violations": [
|
||||
{
|
||||
"type": "duplicate_id",
|
||||
"severity": "error",
|
||||
"collection_pointer": "/groups",
|
||||
"id_list_key": "source_item_ids",
|
||||
"id": "fact_0012",
|
||||
"occurrences": [
|
||||
{
|
||||
"item_index": 1,
|
||||
"id_index": 1
|
||||
},
|
||||
{
|
||||
"item_index": 11,
|
||||
"id_index": 0
|
||||
}
|
||||
],
|
||||
"keep_occurrence": 0,
|
||||
"message": "Source item ID appears in multiple groups."
|
||||
}
|
||||
]
|
||||
}
|
||||
Reference in New Issue
Block a user