mmmikolajczak commited on
Commit
e5fb7ae
·
verified ·
1 Parent(s): 1130a5f

Publish StockNet ACL18 artifact payload

Browse files
.gitattributes CHANGED
@@ -83,3 +83,7 @@ datasets/bigdata22/official/v0/valid.jsonl filter=lfs diff=lfs merge=lfs -text
83
  datasets/bigdata22/official/v0/test.jsonl filter=lfs diff=lfs merge=lfs -text
84
  tasks/bigdata22_stock_movement_v0/requests/train_requests.jsonl filter=lfs diff=lfs merge=lfs -text
85
  tasks/bigdata22_stock_movement_v0/requests/test_requests.jsonl filter=lfs diff=lfs merge=lfs -text
 
 
 
 
 
83
  datasets/bigdata22/official/v0/test.jsonl filter=lfs diff=lfs merge=lfs -text
84
  tasks/bigdata22_stock_movement_v0/requests/train_requests.jsonl filter=lfs diff=lfs merge=lfs -text
85
  tasks/bigdata22_stock_movement_v0/requests/test_requests.jsonl filter=lfs diff=lfs merge=lfs -text
86
+ datasets/stocknet/acl18_paper/v0/test.jsonl filter=lfs diff=lfs merge=lfs -text
87
+ datasets/stocknet/acl18_paper/v0/train.jsonl filter=lfs diff=lfs merge=lfs -text
88
+ tasks/stocknet_acl18_stockmove_v0/requests/test_requests.jsonl filter=lfs diff=lfs merge=lfs -text
89
+ tasks/stocknet_acl18_stockmove_v0/requests/train_requests.jsonl filter=lfs diff=lfs merge=lfs -text
datasets/stocknet/acl18_paper/v0/dev.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
datasets/stocknet/acl18_paper/v0/download_meta.json ADDED
@@ -0,0 +1,73 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "source_repo": "yumoxu/stocknet-dataset",
3
+ "source_url": "https://github.com/yumoxu/stocknet-dataset",
4
+ "source_commit": "330708b5ddc359961078bef469f43f48992fd6e4",
5
+ "source_snapshot_datetime_utc": "2026-04-16T17:26:07.549375+00:00",
6
+ "acquisition_method": "git_clone_checkout_copytree",
7
+ "source_snapshot_root": "data/stocknet_acl18_paper/raw/source_snapshot",
8
+ "source_snapshot_file_count": 55141,
9
+ "source_snapshot_file_sample": [
10
+ "LICENSE",
11
+ "README.md",
12
+ "StockTable",
13
+ "appendix_table_of_target_stocks.pdf",
14
+ "price/preprocessed/AAPL.txt",
15
+ "price/preprocessed/ABB.txt",
16
+ "price/preprocessed/ABBV.txt",
17
+ "price/preprocessed/AEP.txt",
18
+ "price/preprocessed/AGFS.txt",
19
+ "price/preprocessed/AMGN.txt",
20
+ "price/preprocessed/AMZN.txt",
21
+ "price/preprocessed/BA.txt",
22
+ "price/preprocessed/BABA.txt",
23
+ "price/preprocessed/BAC.txt",
24
+ "price/preprocessed/BBL.txt",
25
+ "price/preprocessed/BCH.txt",
26
+ "price/preprocessed/BHP.txt",
27
+ "price/preprocessed/BP.txt",
28
+ "price/preprocessed/BRK-A.txt",
29
+ "price/preprocessed/BSAC.txt",
30
+ "price/preprocessed/BUD.txt",
31
+ "price/preprocessed/C.txt",
32
+ "price/preprocessed/CAT.txt",
33
+ "price/preprocessed/CELG.txt",
34
+ "price/preprocessed/CHL.txt",
35
+ "price/preprocessed/CHTR.txt",
36
+ "price/preprocessed/CMCSA.txt",
37
+ "price/preprocessed/CODI.txt",
38
+ "price/preprocessed/CSCO.txt",
39
+ "price/preprocessed/CVX.txt",
40
+ "price/preprocessed/D.txt",
41
+ "price/preprocessed/DHR.txt",
42
+ "price/preprocessed/DIS.txt",
43
+ "price/preprocessed/DUK.txt",
44
+ "price/preprocessed/EXC.txt",
45
+ "price/preprocessed/FB.txt",
46
+ "price/preprocessed/GD.txt",
47
+ "price/preprocessed/GE.txt",
48
+ "price/preprocessed/GMRE.txt",
49
+ "price/preprocessed/GOOG.txt",
50
+ "price/preprocessed/HD.txt",
51
+ "price/preprocessed/HON.txt",
52
+ "price/preprocessed/HRG.txt",
53
+ "price/preprocessed/HSBC.txt",
54
+ "price/preprocessed/IEP.txt",
55
+ "price/preprocessed/INTC.txt",
56
+ "price/preprocessed/JNJ.txt",
57
+ "price/preprocessed/JPM.txt",
58
+ "price/preprocessed/KO.txt",
59
+ "price/preprocessed/LMT.txt"
60
+ ],
61
+ "source_snapshot_checksums_manifest": "data/stocknet_acl18_paper/raw/source_snapshot_checksums.sha256",
62
+ "sha256_by_core_file": {
63
+ "README.md": "e3f2dee0b0214632a7ab257ffb35396170250a37c2b83077ffcfc5533dfafbdd",
64
+ "LICENSE": "c0fc08ddd92c2e442f5e9fe583c0905e78dc5bda9f5344c376681a0e7e179e51",
65
+ "StockTable": "7dc3b5109b1b702fc8d2d36a2a463d62cc44081964d0060b6f8eeda0943dc1ec",
66
+ "appendix_table_of_target_stocks.pdf": "eabb7aede03dc0ca571b98afc3c095fdf2886a13b96415f7213526b88433c245"
67
+ },
68
+ "notes": [
69
+ "The official StockNet repository is MIT-licensed.",
70
+ "Original source repository is MIT-licensed. The released corpus includes tweet-derived content collected under Twitter's official license and price data sourced from Yahoo Finance; downstream users should review applicable platform/source terms before reuse.",
71
+ "Exact paper counts are reproducible from the official StockNet release when preserving the paper's split boundaries, 5-day alignment, and asymmetric thresholding. A stricter hard tweet-presence filter would drop the release to 18,695 samples across every official dataset snapshot, so ST312 treats that line in the paper as a documented source inconsistency rather than a canonical retention rule."
72
+ ]
73
+ }
datasets/stocknet/acl18_paper/v0/ingest_summary.json ADDED
@@ -0,0 +1,50 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset_id": "stocknet_acl18_paper_v0",
3
+ "canonical_source": "yumoxu/stocknet-dataset",
4
+ "source_commit": "330708b5ddc359961078bef469f43f48992fd6e4",
5
+ "paper_expected_counts": {
6
+ "train": 20339,
7
+ "dev": 2555,
8
+ "test": 3720
9
+ },
10
+ "paper_expected_total": 26614,
11
+ "reconstructed_counts": {
12
+ "train": 20339,
13
+ "dev": 2555,
14
+ "test": 3720
15
+ },
16
+ "reconstructed_total": 26614,
17
+ "number_of_stocks_in_price_release": 88,
18
+ "number_of_stocks_in_tweet_release": 87,
19
+ "number_of_stocks_with_reconstructed_samples": 87,
20
+ "number_of_candidate_stock_days_before_thresholding": 43442,
21
+ "number_removed_by_middle_band_filter": 16820,
22
+ "number_removed_for_insufficient_history": 8,
23
+ "strict_tweet_presence_filter_drop_count": 7924,
24
+ "strict_tweet_presence_filter_counts": {
25
+ "train": 14187,
26
+ "dev": 1871,
27
+ "test": 2632
28
+ },
29
+ "split_counts": {
30
+ "train": 20339,
31
+ "dev": 2555,
32
+ "test": 3720
33
+ },
34
+ "all_paper_count_invariants_passed": true,
35
+ "strict_tweet_presence_filter_matches_paper_counts": false,
36
+ "per_stock_anomalies": {
37
+ "missing_tweet_directories": [
38
+ "GMRE"
39
+ ],
40
+ "stocks_without_reconstructed_samples": [
41
+ "GMRE"
42
+ ]
43
+ },
44
+ "notes": [
45
+ "Canonical reconstruction preserves the paper's exact split dates and total sample counts.",
46
+ "Middle-band filtering uses the paper thresholds <= -0.5% -> Fall and > 0.55% -> Rise.",
47
+ "The official release contains many threshold-retained samples with zero tweets in the 5-day lag; ST312 documents this but does not apply it as a hard filter because doing so breaks the exact paper counts.",
48
+ "Original source repository is MIT-licensed. The released corpus includes tweet-derived content collected under Twitter's official license and price data sourced from Yahoo Finance; downstream users should review applicable platform/source terms before reuse."
49
+ ]
50
+ }
datasets/stocknet/acl18_paper/v0/test.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4a2b36ddba69d73cec1c269a50246ac76e59e0161496e2be3efcba14a3a2dbde
3
+ size 13890777
datasets/stocknet/acl18_paper/v0/train.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:58cd8e013541c8e062bd06805f652f01fecf27ed5872e66a48d639bc7a5292d5
3
+ size 73982295
reports/stocknet_acl18_paper/finben_wrapper_audit.json ADDED
@@ -0,0 +1,126 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "wrapper_dataset": "TheFinAI/flare-sm-acl",
3
+ "wrapper_split_counts_observed": {
4
+ "train": 20781,
5
+ "test": 3720,
6
+ "valid": 2555
7
+ },
8
+ "canonical_split_counts": {
9
+ "train": 20339,
10
+ "dev": 2555,
11
+ "test": 3720
12
+ },
13
+ "field_shape_differences": {
14
+ "wrapper_fields": {
15
+ "train": [
16
+ "id",
17
+ "query",
18
+ "answer",
19
+ "text",
20
+ "choices",
21
+ "gold"
22
+ ],
23
+ "test": [
24
+ "id",
25
+ "query",
26
+ "answer",
27
+ "text",
28
+ "choices",
29
+ "gold"
30
+ ],
31
+ "valid": [
32
+ "id",
33
+ "query",
34
+ "answer",
35
+ "text",
36
+ "choices",
37
+ "gold"
38
+ ]
39
+ },
40
+ "canonical_row_shape": [
41
+ "example_id",
42
+ "dataset_id",
43
+ "split",
44
+ "stock_symbol",
45
+ "target_date",
46
+ "label_int",
47
+ "label_text",
48
+ "target_movement_percent",
49
+ "target_price_prev_adj_close",
50
+ "target_price_adj_close",
51
+ "calendar_lag_days",
52
+ "aligned_trading_days_count",
53
+ "aligned_days",
54
+ "auxiliary_targets",
55
+ "source_dataset",
56
+ "source_policy"
57
+ ]
58
+ },
59
+ "sample_rows": {
60
+ "train": {
61
+ "id": "aclsm0",
62
+ "query": "Scrutinize the data and tweets to envisage if the closing price of $csco will swell or contract at 2014-01-02. Please affirm either Rise or Fall.\nContext: date,open,high,low,close,adj-close,inc-5,inc-10,inc-15,inc-20,inc-25,inc-30\n2013-12-17,-0.9,1.0,-1.1,1.2,1.2,-1.3,-0.0,0.4,0.9,2.1,3.6\n2013-12-18,-0.6,0.2,-2.3,0.4,0.4,-1.6,-0.5,-0.0,0.4,1.2,2.9\n2013-12-19,0.0,0.0,-1.4,0.3,0.3,-1.4,-0.8,-0.4,0.0,0.3,2.2\n2013-12-20,-0.3,0.6,-0.8,0.3,0.3,-0.8,-1.2,-0.7,-0.4,-0.0,1.6\n2013-12-23,-1.5,0.3,-1.6,2.1,2.1,-2.0,-3.0,-2.6,-2.4,-2.1,-0.8\n2013-12-24,-1.1,0.3,-1.2,0.6,0.6,-1.8,-3.3,-3.0,-2.8,-2.5,-1.6\n2013-12-26,-0.4,0.4,-0.4,0.5,0.5,-1.6,-3.4,-3.3,-3.2,-3.0,-2.4\n2013-12-27,-1.0,0.1,-1.1,1.0,1.0,-1.7,-3.7,-4.0,-4.0,-3.8,-3.6\n2013-12-30,-1.1,0.4,-1.1,1.0,1.0,-1.7,-3.8,-4.7,-4.7,-4.6,-4.5\n2013-12-31,-0.8,0.2,-1.2,0.8,0.8,-1.7,-3.8,-5.1,-5.2,-5.2,-5.1\n\n2014-01-01: weekly s&amp;p500 #stocks trend $btu $dhi $jcp $mat $joy $oke $isrg $omc $jnpr $dis $csco $ctxs $se $emn $nbr $fls @ |2013 wrap-up and trading set review - part iii $aapl #apple $bp $cnw $csco $csx $\nAnswer:",
63
+ "answer": "Fall",
64
+ "text": "date,open,high,low,close,adj-close,inc-5,inc-10,inc-15,inc-20,inc-25,inc-30\n2013-12-17,-0.9,1.0,-1.1,1.2,1.2,-1.3,-0.0,0.4,0.9,2.1,3.6\n2013-12-18,-0.6,0.2,-2.3,0.4,0.4,-1.6,-0.5,-0.0,0.4,1.2,2.9\n2013-12-19,0.0,0.0,-1.4,0.3,0.3,-1.4,-0.8,-0.4,0.0,0.3,2.2\n2013-12-20,-0.3,0.6,-0.8,0.3,0.3,-0.8,-1.2,-0.7,-0.4,-0.0,1.6\n2013-12-23,-1.5,0.3,-1.6,2.1,2.1,-2.0,-3.0,-2.6,-2.4,-2.1,-0.8\n2013-12-24,-1.1,0.3,-1.2,0.6,0.6,-1.8,-3.3,-3.0,-2.8,-2.5,-1.6\n2013-12-26,-0.4,0.4,-0.4,0.5,0.5,-1.6,-3.4,-3.3,-3.2,-3.0,-2.4\n2013-12-27,-1.0,0.1,-1.1,1.0,1.0,-1.7,-3.7,-4.0,-4.0,-3.8,-3.6\n2013-12-30,-1.1,0.4,-1.1,1.0,1.0,-1.7,-3.8,-4.7,-4.7,-4.6,-4.5\n2013-12-31,-0.8,0.2,-1.2,0.8,0.8,-1.7,-3.8,-5.1,-5.2,-5.2,-5.1\n\n2014-01-01: weekly s&amp;p500 #stocks trend $btu $dhi $jcp $mat $joy $oke $isrg $omc $jnpr $dis $csco $ctxs $se $emn $nbr $fls @ |2013 wrap-up and trading set review - part iii $aapl #apple $bp $cnw $csco $csx $",
65
+ "choices": [
66
+ "Rise",
67
+ "Fall"
68
+ ],
69
+ "gold": 1
70
+ },
71
+ "test": {
72
+ "id": "aclsm23336",
73
+ "query": "By reviewing the data and tweets, can we predict if the closing price of $csco will go upwards or downwards at 2015-10-01? Please indicate either Rise or Fall.\nContext: date,open,high,low,close,adj-close,inc-5,inc-10,inc-15,inc-20,inc-25,inc-30\n2015-09-17,0.3,1.5,-0.3,-0.7,-0.7,0.2,0.3,0.0,-0.0,2.0,3.2\n2015-09-18,0.1,0.8,-1.1,-1.3,-1.3,1.2,1.5,1.2,1.0,2.9,4.2\n2015-09-21,-0.4,1.1,-0.7,0.0,0.0,1.0,1.5,1.1,0.8,2.3,3.8\n2015-09-22,-0.3,0.3,-0.7,-1.6,-1.6,2.0,2.7,2.5,2.4,3.4,5.0\n2015-09-23,-0.2,0.1,-1.1,0.6,0.6,0.8,1.8,2.0,2.0,2.3,4.1\n2015-09-24,-1.3,0.4,-1.9,0.5,0.5,-0.1,1.0,1.4,1.4,1.4,3.2\n2015-09-25,-0.8,1.0,-0.9,2.4,2.4,-2.1,-1.4,-1.0,-1.0,-1.1,0.4\n2015-09-28,0.7,1.4,-0.3,-1.4,-1.4,-0.6,-0.0,0.5,0.3,0.2,1.4\n2015-09-29,0.0,1.1,-0.3,-0.1,-0.1,-0.1,-0.1,0.4,0.3,0.3,1.1\n2015-09-30,-1.0,0.5,-1.4,2.4,2.4,-1.7,-2.3,-1.9,-1.8,-1.8,-1.5\n\n2015-09-19: csco cisco systems, inc. p/e ratio$csco $cprt $tanh $eca.to #csco #investing #stock|csco cisco systems, inc. short sales$csco $gs $dbb $pfe #csco #pennystocks #pennystocks\n2015-09-23: gdxj market vectors junior gold miners volume$gdxj $xly $uup $csco #gdxj #pennystocks #invest|cisco helped build china's internet, now struggles push it to strike a deal. $csco via AT_USER|$spy bull \n2015-09-24: will cisco systems, inc. $csco will be able to crack a deal in china? |bnd vanguard total bond market ask$bnd $csco $gdx $joy #bnd #tradeideas #stock\n2015-09-25: saluting the other great #norcal #tech cos including: $aapl $goog $tsla $crm $fb $intc $twtr $adbe $adsk $csco $nflx $ebay $dna $hpq $orcl.|$fb $msft $adbe $twtr $csco $qcom:which high-tech heavy hitt\n2015-09-26: rt rt cvc cablevision systems investor opinions$cvc $ewi $csco…\n2015-09-27: apdn applied dna sciences inc. tick$apdn $csco $tvix $infn #apdn #nasdaq #invest|qld proshares ultra qqq fund short sales$qld $aav.to $csco $iwm #qld #nasdaq #investingqtmm quantum materials…\n2015-09-28: csco cisco systems, inc. stock message board$csco $jjc $tsla $t #csco #pennystocks #nasdaqcsh/un.to chartwell…|$met news: \"i walk the line $cmcsa $csco $cvc $ebay $emc $gm $met $pypl $sbgi …\" analyz\n2015-09-29: fro frontline ltd. company info$fro $avxl $ziop $csco #fro #tradeideas #stockmarket|csco cisco systems, inc. shares$csco $gild $oil $spxu #csco #tradeideas #share|dis walt disney co. eod data$dis $hzn\n2015-09-30: our platinum members have been seeing massive gains on our penny stock picks! subscribe today: $amzn $qcom $csco|$csco - proxy statement (definitive) (def 14a) |recent buy: cisco systems $csco\nAnswer:",
74
+ "answer": "Fall",
75
+ "text": "date,open,high,low,close,adj-close,inc-5,inc-10,inc-15,inc-20,inc-25,inc-30\n2015-09-17,0.3,1.5,-0.3,-0.7,-0.7,0.2,0.3,0.0,-0.0,2.0,3.2\n2015-09-18,0.1,0.8,-1.1,-1.3,-1.3,1.2,1.5,1.2,1.0,2.9,4.2\n2015-09-21,-0.4,1.1,-0.7,0.0,0.0,1.0,1.5,1.1,0.8,2.3,3.8\n2015-09-22,-0.3,0.3,-0.7,-1.6,-1.6,2.0,2.7,2.5,2.4,3.4,5.0\n2015-09-23,-0.2,0.1,-1.1,0.6,0.6,0.8,1.8,2.0,2.0,2.3,4.1\n2015-09-24,-1.3,0.4,-1.9,0.5,0.5,-0.1,1.0,1.4,1.4,1.4,3.2\n2015-09-25,-0.8,1.0,-0.9,2.4,2.4,-2.1,-1.4,-1.0,-1.0,-1.1,0.4\n2015-09-28,0.7,1.4,-0.3,-1.4,-1.4,-0.6,-0.0,0.5,0.3,0.2,1.4\n2015-09-29,0.0,1.1,-0.3,-0.1,-0.1,-0.1,-0.1,0.4,0.3,0.3,1.1\n2015-09-30,-1.0,0.5,-1.4,2.4,2.4,-1.7,-2.3,-1.9,-1.8,-1.8,-1.5\n\n2015-09-19: csco cisco systems, inc. p/e ratio$csco $cprt $tanh $eca.to #csco #investing #stock|csco cisco systems, inc. short sales$csco $gs $dbb $pfe #csco #pennystocks #pennystocks\n2015-09-23: gdxj market vectors junior gold miners volume$gdxj $xly $uup $csco #gdxj #pennystocks #invest|cisco helped build china's internet, now struggles push it to strike a deal. $csco via AT_USER|$spy bull \n2015-09-24: will cisco systems, inc. $csco will be able to crack a deal in china? |bnd vanguard total bond market ask$bnd $csco $gdx $joy #bnd #tradeideas #stock\n2015-09-25: saluting the other great #norcal #tech cos including: $aapl $goog $tsla $crm $fb $intc $twtr $adbe $adsk $csco $nflx $ebay $dna $hpq $orcl.|$fb $msft $adbe $twtr $csco $qcom:which high-tech heavy hitt\n2015-09-26: rt rt cvc cablevision systems investor opinions$cvc $ewi $csco…\n2015-09-27: apdn applied dna sciences inc. tick$apdn $csco $tvix $infn #apdn #nasdaq #invest|qld proshares ultra qqq fund short sales$qld $aav.to $csco $iwm #qld #nasdaq #investingqtmm quantum materials…\n2015-09-28: csco cisco systems, inc. stock message board$csco $jjc $tsla $t #csco #pennystocks #nasdaqcsh/un.to chartwell…|$met news: \"i walk the line $cmcsa $csco $cvc $ebay $emc $gm $met $pypl $sbgi …\" analyz\n2015-09-29: fro frontline ltd. company info$fro $avxl $ziop $csco #fro #tradeideas #stockmarket|csco cisco systems, inc. shares$csco $gild $oil $spxu #csco #tradeideas #share|dis walt disney co. eod data$dis $hzn\n2015-09-30: our platinum members have been seeing massive gains on our penny stock picks! subscribe today: $amzn $qcom $csco|$csco - proxy statement (definitive) (def 14a) |recent buy: cisco systems $csco",
76
+ "choices": [
77
+ "Rise",
78
+ "Fall"
79
+ ],
80
+ "gold": 1
81
+ },
82
+ "valid": {
83
+ "id": "aclsm20781",
84
+ "query": "By reviewing the data and tweets, can we predict if the closing price of $v will go upwards or downwards at 2015-08-03? Please indicate either Rise or Fall.\nContext: date,open,high,low,close,adj-close,inc-5,inc-10,inc-15,inc-20,inc-25,inc-30\n2015-07-20,-1.9,0.4,-2.0,2.6,2.6,-2.5,-4.6,-5.4,-5.3,-5.3,-5.3\n2015-07-21,1.1,1.2,-0.4,-0.9,-0.9,-1.1,-3.1,-4.0,-4.2,-4.2,-4.2\n2015-07-22,-0.0,0.3,-0.5,-0.1,-0.1,-0.5,-2.3,-3.5,-4.0,-4.0,-4.0\n2015-07-23,0.6,0.7,-0.8,-0.3,-0.3,0.2,-1.3,-2.8,-3.5,-3.5,-3.6\n2015-07-24,2.6,2.8,-0.5,4.3,4.3,-2.9,-4.5,-6.2,-7.0,-7.2,-7.3\n2015-07-27,0.6,1.2,-0.4,-1.3,-1.3,-1.3,-2.7,-4.5,-5.5,-5.7,-5.9\n2015-07-28,-0.6,0.6,-1.2,1.2,1.2,-1.8,-3.2,-5.0,-6.1,-6.5,-6.7\n2015-07-29,-1.4,0.6,-1.8,1.7,1.7,-2.3,-4.1,-5.8,-7.0,-7.7,-8.0\n2015-07-30,-0.6,0.5,-1.5,0.5,0.5,-1.6,-3.8,-5.4,-6.9,-7.8,-8.1\n2015-07-31,1.5,1.6,-0.0,-1.4,-1.4,-0.1,-1.8,-3.5,-5.2,-6.2,-6.6\n\n2015-07-20: $gs $v $erx $xiv $f july 20th, 2015 watch list - update from option millionaires - stock options chatroom futures... \n2015-07-21: #wallstreet: #dowjones +0.08%, en 18,100.41, #sp500 +0.08% en 2,128.28 y #nasdaq +0.17%. mejor valor: $v +2.57% pe… |dbc db commodities tracking index fund 52wk low$dbc $v $jbl $ng #dbc #invest #stock\n2015-07-22: AT_USER $v visa declares $0.12 dividend |my rental property calculator $bac $usb $wfc $key $jpm $v $ma $dfs $axp|v visa inc. stock forum$v $googl $mu $abc #v #investing #nasdaq|$gs $v $erx $xiv $f we\n2015-07-23: ijjp ijj corp. shares$ijjp $v $luv $xeg.to #ijjp #share #shareinfn infinera corp. dividend…|AT_USER $v just reported earnings for q3 2015 – |rt AT_USER $amzn $v $sbux all up big a/h...strong stocks ge\n2015-07-24: rt AT_USER $v's charlie scharf told it like it is v. #fifa--totally unacceptable behavior. watch #fifa change tomorrow. $v is too power…|fv currency shares canadian dollar trust ex-div date$fv $uso $v\n2015-07-25: v visa inc. last trade$v $gdx $xlv $goog #v #investing #tradeideas|$v message board : \n2015-07-27: \"redrum\" monday? $spy $spx $qqq $iwm $djia $dji $es_f $googl $amzn $aapl $v $p $twtr $fb #bolsa #stocks #trading … |$cat $dd $utx $v trading july's biggest movers month to date \n2015-07-29: ocls oculus innovative sciences inc. investor opinions$ocls $v $iyr $xbi #ocls #stockmarket #shareoih market…|bwld buffalo wild wings inc. message board posts$bwld $ung $jnj $v #bwld #pennystocks #nas\n2015-07-30: vrx valeant pharmaceuticals intl, inc. 52wk high$vrx $rsp $vxx $v #vrx #investing #stockmarket|jake132013 : mnkd mannkind corp. ask size$mnkd $xlp $gdx $v #mn… |fbiox fidelity select biotechnology la\n2015-07-31: v visa inc. chart$v $uso $ew $agn #v #nasdaq #share|*g=tfpxcpte$v&lt;t)$fd:v^p6;oej4r?8lpj?\n2015-08-01: fxy currency shares japanese yen trust p/e ratio$fxy $rig $v $goog #fxy #stockmarket #investing|tsla tesla motors inc. filings$tsla $mdy $spxu $v #tsla #nasdaq #tradeideas|gmcr keurig green mountain, \n2015-08-02: $gs $v $erx $xiv $f the week ahead – august 2nd, 2015 - update from option millionaires - stock options chatroom ... |v visa inc. tick$v $ibm $epi $agn #v #finance #stock|v visa inc. open$v $tfi.to $g\nAnswer:",
85
+ "answer": "Rise",
86
+ "text": "date,open,high,low,close,adj-close,inc-5,inc-10,inc-15,inc-20,inc-25,inc-30\n2015-07-20,-1.9,0.4,-2.0,2.6,2.6,-2.5,-4.6,-5.4,-5.3,-5.3,-5.3\n2015-07-21,1.1,1.2,-0.4,-0.9,-0.9,-1.1,-3.1,-4.0,-4.2,-4.2,-4.2\n2015-07-22,-0.0,0.3,-0.5,-0.1,-0.1,-0.5,-2.3,-3.5,-4.0,-4.0,-4.0\n2015-07-23,0.6,0.7,-0.8,-0.3,-0.3,0.2,-1.3,-2.8,-3.5,-3.5,-3.6\n2015-07-24,2.6,2.8,-0.5,4.3,4.3,-2.9,-4.5,-6.2,-7.0,-7.2,-7.3\n2015-07-27,0.6,1.2,-0.4,-1.3,-1.3,-1.3,-2.7,-4.5,-5.5,-5.7,-5.9\n2015-07-28,-0.6,0.6,-1.2,1.2,1.2,-1.8,-3.2,-5.0,-6.1,-6.5,-6.7\n2015-07-29,-1.4,0.6,-1.8,1.7,1.7,-2.3,-4.1,-5.8,-7.0,-7.7,-8.0\n2015-07-30,-0.6,0.5,-1.5,0.5,0.5,-1.6,-3.8,-5.4,-6.9,-7.8,-8.1\n2015-07-31,1.5,1.6,-0.0,-1.4,-1.4,-0.1,-1.8,-3.5,-5.2,-6.2,-6.6\n\n2015-07-20: $gs $v $erx $xiv $f july 20th, 2015 watch list - update from option millionaires - stock options chatroom futures... \n2015-07-21: #wallstreet: #dowjones +0.08%, en 18,100.41, #sp500 +0.08% en 2,128.28 y #nasdaq +0.17%. mejor valor: $v +2.57% pe… |dbc db commodities tracking index fund 52wk low$dbc $v $jbl $ng #dbc #invest #stock\n2015-07-22: AT_USER $v visa declares $0.12 dividend |my rental property calculator $bac $usb $wfc $key $jpm $v $ma $dfs $axp|v visa inc. stock forum$v $googl $mu $abc #v #investing #nasdaq|$gs $v $erx $xiv $f we\n2015-07-23: ijjp ijj corp. shares$ijjp $v $luv $xeg.to #ijjp #share #shareinfn infinera corp. dividend…|AT_USER $v just reported earnings for q3 2015 – |rt AT_USER $amzn $v $sbux all up big a/h...strong stocks ge\n2015-07-24: rt AT_USER $v's charlie scharf told it like it is v. #fifa--totally unacceptable behavior. watch #fifa change tomorrow. $v is too power…|fv currency shares canadian dollar trust ex-div date$fv $uso $v\n2015-07-25: v visa inc. last trade$v $gdx $xlv $goog #v #investing #tradeideas|$v message board : \n2015-07-27: \"redrum\" monday? $spy $spx $qqq $iwm $djia $dji $es_f $googl $amzn $aapl $v $p $twtr $fb #bolsa #stocks #trading … |$cat $dd $utx $v trading july's biggest movers month to date \n2015-07-29: ocls oculus innovative sciences inc. investor opinions$ocls $v $iyr $xbi #ocls #stockmarket #shareoih market…|bwld buffalo wild wings inc. message board posts$bwld $ung $jnj $v #bwld #pennystocks #nas\n2015-07-30: vrx valeant pharmaceuticals intl, inc. 52wk high$vrx $rsp $vxx $v #vrx #investing #stockmarket|jake132013 : mnkd mannkind corp. ask size$mnkd $xlp $gdx $v #mn… |fbiox fidelity select biotechnology la\n2015-07-31: v visa inc. chart$v $uso $ew $agn #v #nasdaq #share|*g=tfpxcpte$v&lt;t)$fd:v^p6;oej4r?8lpj?\n2015-08-01: fxy currency shares japanese yen trust p/e ratio$fxy $rig $v $goog #fxy #stockmarket #investing|tsla tesla motors inc. filings$tsla $mdy $spxu $v #tsla #nasdaq #tradeideas|gmcr keurig green mountain, \n2015-08-02: $gs $v $erx $xiv $f the week ahead – august 2nd, 2015 - update from option millionaires - stock options chatroom ... |v visa inc. tick$v $ibm $epi $agn #v #finance #stock|v visa inc. open$v $tfi.to $g",
87
+ "choices": [
88
+ "Rise",
89
+ "Fall"
90
+ ],
91
+ "gold": 0
92
+ }
93
+ },
94
+ "threshold_rule_difference_note": "The FLARE wrapper is promptized and aligns with a simplified public description, while the canonical ST312 module preserves the paper-level asymmetric thresholds <= -0.5% / > 0.55% and exact ACL18 split counts.",
95
+ "wrapper_date_range_and_lag_window_note": "Wrapper examples appear to present a 10-row flattened historical context and observed split counts 20,781 / 2,555 / 3,720, which do not match the paper's 20,339 / 2,555 / 3,720 split counts.",
96
+ "answer_gold_semantics": {
97
+ "train": {
98
+ "answer_label": "Fall",
99
+ "gold_label": "Fall",
100
+ "choices": [
101
+ "Rise",
102
+ "Fall"
103
+ ]
104
+ },
105
+ "test": {
106
+ "answer_label": "Fall",
107
+ "gold_label": "Fall",
108
+ "choices": [
109
+ "Rise",
110
+ "Fall"
111
+ ]
112
+ },
113
+ "valid": {
114
+ "answer_label": "Rise",
115
+ "gold_label": "Rise",
116
+ "choices": [
117
+ "Rise",
118
+ "Fall"
119
+ ]
120
+ }
121
+ },
122
+ "notes": [
123
+ "The FLARE wrapper is audited only as a compatibility surface and does not define the canonical StockNet dataset.",
124
+ "Wrapper answer values appear to use Rise/Fall directly, while wrapper gold values appear to be choice indices."
125
+ ]
126
+ }
reports/stocknet_acl18_paper/raw_schema_summary.json ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset_id": "stocknet_acl18_paper_v0",
3
+ "source_repo": "yumoxu/stocknet-dataset",
4
+ "source_commit": "330708b5ddc359961078bef469f43f48992fd6e4",
5
+ "price_preprocessed_file_count": 88,
6
+ "price_raw_file_count": 88,
7
+ "tweet_preprocessed_dir_count": 87,
8
+ "tweet_raw_dir_count": 87,
9
+ "missing_tweet_dirs": [
10
+ "GMRE"
11
+ ],
12
+ "sample_price_preprocessed_fields": [
13
+ "date",
14
+ "movement_percent",
15
+ "open_norm",
16
+ "high_norm",
17
+ "low_norm",
18
+ "close_norm",
19
+ "volume"
20
+ ],
21
+ "sample_price_raw_fields": [
22
+ "date",
23
+ "open",
24
+ "high",
25
+ "low",
26
+ "close",
27
+ "adj_close",
28
+ "volume"
29
+ ],
30
+ "sample_preprocessed_tweet_texts": [
31
+ "dividend portfolio playoffs wild card game : wells fargo vs . dupont URL $ wfc $ dd"
32
+ ],
33
+ "notes": [
34
+ "The official StockNet release exposes 88 price files and 87 tweet directories at the pinned source snapshot.",
35
+ "GMRE appears in the price release but has no corresponding tweet directory and no canonical paper-window samples.",
36
+ "Preprocessed tweet texts are token lists that ST312 joins back into space-separated strings for canonical JSONL serialization.",
37
+ "Price movement percents in the preprocessed release match adjusted-close returns computed from the raw Yahoo Finance CSVs within float tolerance."
38
+ ]
39
+ }
reports/stocknet_acl18_paper/reconstruction_audit.json ADDED
@@ -0,0 +1,237 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "status": "success",
3
+ "paper_expected_counts": {
4
+ "train": 20339,
5
+ "dev": 2555,
6
+ "test": 3720
7
+ },
8
+ "paper_expected_total": 26614,
9
+ "reconstructed_counts": {
10
+ "train": 20339,
11
+ "dev": 2555,
12
+ "test": 3720
13
+ },
14
+ "reconstructed_total": 26614,
15
+ "candidate_counts_before_filtering": {
16
+ "train": 34220,
17
+ "dev": 3654,
18
+ "test": 5568
19
+ },
20
+ "removal_counts": {
21
+ "middle_band": 16820,
22
+ "insufficient_history": 8
23
+ },
24
+ "strict_tweet_presence_filter_counts": {
25
+ "train": 14187,
26
+ "dev": 1871,
27
+ "test": 2632
28
+ },
29
+ "strict_tweet_presence_filter_total": 18690,
30
+ "strict_tweet_presence_matches_paper_counts": false,
31
+ "strict_tweet_presence_drop_count": 7924,
32
+ "source_snapshot_commit": "330708b5ddc359961078bef469f43f48992fd6e4",
33
+ "stocks_in_price_release": 88,
34
+ "stocks_in_tweet_release": 87,
35
+ "stocks_with_reconstructed_samples": 87,
36
+ "per_stock_sample_counts": {
37
+ "AAPL": 336,
38
+ "ABB": 307,
39
+ "ABBV": 351,
40
+ "AEP": 306,
41
+ "AGFS": 147,
42
+ "AMGN": 354,
43
+ "AMZN": 348,
44
+ "BA": 322,
45
+ "BABA": 237,
46
+ "BAC": 336,
47
+ "BBL": 382,
48
+ "BCH": 313,
49
+ "BHP": 379,
50
+ "BP": 315,
51
+ "BRK-A": 253,
52
+ "BSAC": 341,
53
+ "BUD": 317,
54
+ "C": 335,
55
+ "CAT": 334,
56
+ "CELG": 377,
57
+ "CHL": 330,
58
+ "CHTR": 337,
59
+ "CMCSA": 310,
60
+ "CODI": 306,
61
+ "CSCO": 310,
62
+ "CVX": 312,
63
+ "D": 296,
64
+ "DHR": 273,
65
+ "DIS": 299,
66
+ "DUK": 283,
67
+ "EXC": 323,
68
+ "FB": 365,
69
+ "GD": 297,
70
+ "GE": 296,
71
+ "GOOG": 320,
72
+ "HD": 284,
73
+ "HON": 274,
74
+ "HRG": 333,
75
+ "HSBC": 296,
76
+ "IEP": 332,
77
+ "INTC": 326,
78
+ "JNJ": 263,
79
+ "JPM": 306,
80
+ "KO": 239,
81
+ "LMT": 295,
82
+ "MA": 315,
83
+ "MCD": 254,
84
+ "MDT": 299,
85
+ "MMM": 257,
86
+ "MO": 273,
87
+ "MRK": 310,
88
+ "MSFT": 327,
89
+ "NEE": 301,
90
+ "NGG": 280,
91
+ "NVS": 280,
92
+ "ORCL": 305,
93
+ "PCG": 317,
94
+ "PCLN": 345,
95
+ "PEP": 258,
96
+ "PFE": 279,
97
+ "PG": 230,
98
+ "PICO": 371,
99
+ "PM": 264,
100
+ "PPL": 321,
101
+ "PTR": 371,
102
+ "RDS-B": 323,
103
+ "REX": 404,
104
+ "SLB": 324,
105
+ "SNP": 364,
106
+ "SNY": 326,
107
+ "SO": 269,
108
+ "SPLP": 239,
109
+ "SRE": 289,
110
+ "T": 266,
111
+ "TM": 292,
112
+ "TOT": 359,
113
+ "TSM": 351,
114
+ "UL": 280,
115
+ "UN": 297,
116
+ "UNH": 319,
117
+ "UPS": 256,
118
+ "UTX": 286,
119
+ "V": 309,
120
+ "VZ": 255,
121
+ "WFC": 290,
122
+ "WMT": 266,
123
+ "XOM": 298
124
+ },
125
+ "per_stock_no_tweet_counts": {
126
+ "AAPL": 3,
127
+ "ABB": 196,
128
+ "ABBV": 25,
129
+ "AEP": 143,
130
+ "AGFS": 141,
131
+ "AMGN": 20,
132
+ "AMZN": 5,
133
+ "BA": 9,
134
+ "BABA": 4,
135
+ "BAC": 1,
136
+ "BBL": 236,
137
+ "BCH": 300,
138
+ "BHP": 85,
139
+ "BP": 31,
140
+ "BRK-A": 235,
141
+ "BSAC": 309,
142
+ "BUD": 149,
143
+ "C": 2,
144
+ "CAT": 16,
145
+ "CELG": 16,
146
+ "CHL": 135,
147
+ "CHTR": 118,
148
+ "CMCSA": 31,
149
+ "CODI": 253,
150
+ "CSCO": 6,
151
+ "CVX": 21,
152
+ "D": 2,
153
+ "DHR": 134,
154
+ "DIS": 6,
155
+ "DUK": 108,
156
+ "EXC": 124,
157
+ "GD": 76,
158
+ "GE": 6,
159
+ "HD": 19,
160
+ "HON": 87,
161
+ "HRG": 273,
162
+ "HSBC": 162,
163
+ "IEP": 204,
164
+ "INTC": 7,
165
+ "JNJ": 8,
166
+ "JPM": 7,
167
+ "KO": 6,
168
+ "LMT": 43,
169
+ "MA": 23,
170
+ "MCD": 7,
171
+ "MDT": 69,
172
+ "MMM": 29,
173
+ "MO": 32,
174
+ "MRK": 13,
175
+ "MSFT": 3,
176
+ "NEE": 117,
177
+ "NGG": 219,
178
+ "NVS": 75,
179
+ "ORCL": 14,
180
+ "PCG": 131,
181
+ "PCLN": 7,
182
+ "PEP": 37,
183
+ "PFE": 7,
184
+ "PG": 15,
185
+ "PICO": 341,
186
+ "PM": 84,
187
+ "PPL": 144,
188
+ "PTR": 287,
189
+ "RDS-B": 315,
190
+ "REX": 265,
191
+ "SLB": 44,
192
+ "SNP": 296,
193
+ "SNY": 107,
194
+ "SO": 68,
195
+ "SPLP": 226,
196
+ "SRE": 171,
197
+ "T": 1,
198
+ "TM": 127,
199
+ "TOT": 164,
200
+ "TSM": 183,
201
+ "UL": 182,
202
+ "UN": 186,
203
+ "UNH": 35,
204
+ "UPS": 53,
205
+ "UTX": 50,
206
+ "V": 7,
207
+ "VZ": 7,
208
+ "WFC": 8,
209
+ "WMT": 6,
210
+ "XOM": 7
211
+ },
212
+ "aligned_trading_days_distribution": {
213
+ "2": 1557,
214
+ "3": 14846,
215
+ "4": 5518,
216
+ "5": 4693
217
+ },
218
+ "class_counts_by_split": {
219
+ "train": {
220
+ "Fall": 10021,
221
+ "Rise": 10318
222
+ },
223
+ "dev": {
224
+ "Fall": 1416,
225
+ "Rise": 1139
226
+ },
227
+ "test": {
228
+ "Fall": 1812,
229
+ "Rise": 1908
230
+ }
231
+ },
232
+ "notes": [
233
+ "Canonical ST312 reconstruction preserves the paper's exact split boundaries, 5-day alignment semantics, and asymmetric thresholds.",
234
+ "Exact paper counts are reproducible from the official StockNet release when preserving the paper's split boundaries, 5-day alignment, and asymmetric thresholding. A stricter hard tweet-presence filter would drop the release to 18,695 samples across every official dataset snapshot, so ST312 treats that line in the paper as a documented source inconsistency rather than a canonical retention rule.",
235
+ "This is an inference from the official dataset release and reference code, because every official source snapshot reproduces the paper counts only without the extra hard tweet-presence filter."
236
+ ]
237
+ }
tasks/stocknet_acl18_stockmove_v0/README.md ADDED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # stocknet_acl18_stockmove_v0
2
+
3
+ Canonical ACL18 StockNet stock movement prediction task over `stocknet_acl18_paper_v0`.
4
+
5
+ ## Formal task id
6
+ `FC_STOCKMOVE_STOCKNET_ACL18_v0`
7
+
8
+ ## Task definition
9
+ - canonical output is JSON with exactly one key: `label`
10
+ - allowed labels are exactly `Rise` or `Fall`
11
+ - canonical task is the paper-level binary stock movement forecasting problem from 5-day aligned tweet corpora plus historical stock prices
12
+ - this is the original ACL18 task, not a copied FLARE prompt wrapper
13
+
14
+ ## Evaluation variants
15
+ - `paper_default`: canonical Accuracy + MCC evaluation over the exact paper test split
16
+ - `finben_acl18_optional`: compatibility parser/view that accepts canonical `label` outputs and optional wrapper-style `answer` outputs while still reporting Accuracy + MCC
17
+
18
+ ## Canonical metrics
19
+ - benchmark metrics: `accuracy`, `mcc`
20
+ - operational diagnostics: `format_valid_rate`, `completion_coverage`, `nonempty_label_rate`
21
+
22
+ ## Notes
23
+ - reward helpers are exploratory and offline-only; canonical model comparison uses the cached evaluator, not the reward totals
24
+ - the dataset README and reconstruction audit document a source inconsistency around tweet-presence filtering versus exact paper counts
25
+ - Original source repository is MIT-licensed. The released corpus includes tweet-derived content collected under Twitter's official license and price data sourced from Yahoo Finance; downstream users should review applicable platform/source terms before reuse.
26
+
27
+ ## Output schema
28
+ Return ONLY a JSON object of this form:
29
+
30
+ ```json
31
+ {
32
+ "label": "Rise"
33
+ }
34
+ ```
tasks/stocknet_acl18_stockmove_v0/requests/dev_requests.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
tasks/stocknet_acl18_stockmove_v0/requests/test_requests.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:69cb287592c9e0f7eac381f5bd8bb4f5b8e96cab8705e2c9c24d0fc5d2c3d90d
3
+ size 11014881
tasks/stocknet_acl18_stockmove_v0/requests/train_requests.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:127682313639058051b2151118434ae2b54902373dc5ab7bf38585f5531c11bd
3
+ size 58472479