Open Hub
Open Hub

 Settings |  Report Duplicate

0
I Use This!
Very Low Activity

Commits : Listings

Analyzed about 8 hours ago. based on code collected 1 day ago.
Nov 05, 2025 — Nov 05, 2026
Commit Message Contributor Files Modified Lines Added Lines Removed Code Location Date
Updated Requirements.txt
Srinivasan Kannan
More... almost 8 years ago
Theory - Balls and Bins Problem and Theoretical EVMs
Srinivasan Kannan
More... almost 8 years ago
Correction - reference - Riemann Zeta Function
Srinivasan Kannan
More... almost 8 years ago
Theory - PACELC Theorem, CAP Theorem and limitations of Electronic Voting Machines on cloud
Srinivasan Kannan
More... almost 8 years ago
Reference - Cloud - CAP Theorem
Srinivasan Kannan
More... almost 8 years ago
Theory and Feature - Software Analytics - Cyclomatic Complexity - FTrace kernel function call graph analysis
Srinivasan Kannan
More... almost 8 years ago
---------------------------------------------------------------------------------------------------------------- 619. (THEORY and FEATURE) Software Analytics - Cyclomatic Complexity - FTrace call graphs - 28 November 2018 - related to 581 ---------------------------------------------------------------------------------------------------------------- 1. FTrace kernel function call graph shell script (asfer_ftrace.sh) already in USBmd and USBmd64 have been included as part of AsFer Cyclomatic Software Analytics for analyzing what happens in kernel within while executing a userspace code. 2. Example ftrace.log for Computational Geometric Factorization has been recorded in ftrace.DiscreteHyperbolicFactorization_TileSearch_Optimized.log 3. DOT graph creation function for parsing ftrace call graph log has been added in CyclomaticComplexitySparkMapReducer.py 4. FTrace call graph DOT file is written to CyclomaticComplexitySparkMapReducer.ftrace_callgraph.dot 5. logs for Cyclomatic Analysis of kernel callgraph for previous factorization example have been committed to: testlogs/CyclomaticComplexitySparkMapReducer.log.FTrace.28November2018 testlogs/DiscreteHyperbolicFactorizationUpperbound_TileSearch_Optimized.factors testlogs/DiscreteHyperbolicFactorizationUpperbound_TileSearch_Optimized.ftrace.log.28November2018 6. PageRank, Degree Centrality, Cycles and Longest Path of Call graph are printed. 7. Least pageranked kernel functions are: ... ('unix_release\n', 0.0006824636787772093), ('nf_ct_get_tuple\n', 0.0006824636787772093), ('insert_work\n', 0.0006824636787772093), ('wake_up_worker\n', 0.0006824636787772093), ('__tcp_push_pending_frames\n', 0.0006824636787772093), ('bictcp_cong_avoid\n', 0.0006824636787772093), ('fsnotify_clear_marks_by_inode\n', 0.0006824636787772093), ('tcp_timewait_state_process\n', 0.0006824636787772093), ('ktime_get\n', 0.0006824636787772093), ('free_page_and_swap_cache\n', 0.0006824636787772093), ('update_page_reclaim_stat\n', 0.0006824636787772093), ('__cond_resched\n', 0.0006824636787772093), ('unlink_anon_vmas\n', 0.0006824636787772093), ('timer_interrupt\n', 0.0006824636787772093), ('inet_put_port\n', 0.0006824636787772093), ('update_process_times\n', 0.0006824636787772093), ('tcp_v4_send_ack\n', 0.0006824636787772093), ('ext4_get_inode_loc\n', 0.0006824636787772093), ('enable_8259A_irq\n', 0.0006824636787772093), ('__pagevec_free\n', 0.0006824636787772093), ('getnstimeofday\n', 0.0006824636787772093), ('dequeue_task\n', 0.0006824636787772093), ('__jbd2_journal_file_buffer\n', 0.0006824636787772093), ('fuse_prepare_release\n', 0.0006824636787772093), ('tcp_established_options\n', 0.0006824636787772093), ('fuse_release_common\n', 0.0006824636787772093) ... which are top level invocations closest to userspace arising out of TCP traffic in Spark Executor Driver (TCP in kernel: http://vger.kernel.org/~davem/tcp_output.html) 8. It is worthy to note lot of ----exit() kernel functions related to VMA paging (e.g.do_exit()) which release virtual memory pages (possibly by GC) and TCP (tcp_write_xmit(), tcp_transmit_skb(), etc.,) having high degree centrality showing heavy TCP traffic.
Srinivasan Kannan
More... almost 8 years ago
Correction - Rectifiable Curves
Srinivasan Kannan
More... almost 8 years ago
Theory - Rectification - Pixelation of Hyperbolic Arc - Computational Geometric Factorization
Srinivasan Kannan
More... almost 8 years ago
Theory - Kummer Theorem for Diophantine Representation of Arbitray RE Set
Srinivasan Kannan
More... almost 8 years ago
Theory - Tiling and Diophantines - ABC Conjecture
Srinivasan Kannan
More... almost 8 years ago
Feature - People Analytics - Connections Analysis
Srinivasan Kannan
More... almost 8 years ago
----------------------------------------------------------------------------------------------------------------- 618. (FEATURE) People Analytics - Experiential Intrinsic Merit - LinkedIn Profile Connections - 20 November 2018 - related to 443 ----------------------------------------------------------------------------------------------------------------- 1.SocialNetworkAnalysis_PeopleAnalytics.py has been changed to parse connections of a linkedin profile 2.Number of connections of a profile is then used to compute experiential intrinsic merit based on degree of a vertex as: E = M*e^(kMt) = log(dv(t)) * e^(klog(dv(t))*t/clogt) / clogt for evolving degree dv(t) 3.This identity is derived from exponential relation between Fame (Degree) and Merit (Intrinsic fitness) of a social profile - in https://www.microsoft.com/en-us/research/wp-content/uploads/2016/11/First-to-Market-is-not-Everything-an-Analysis-of-Preferential-Attachment-with-Fitness.pdf (described earlier) 4.Previous relation is anachronistic in the sense merit is defined in terms of fame when it has to be other way around. Function for experiential_intrinsic_merit() takes number of connections as a parameter and branches off into two clauses for least enery log normal merit if degree is 0 and degree based merit for degree > 0. 5.Example LinkedIn Profile Connections (committed in pdf and txt) have been analyzed for previous degree based experiential merit and Recursive Lambda Function Growth merit. Example linkedin Connection pdf and text file have been committed to: ConnectionsLinkedIn_KSrinivasan.pdf ConnectionsLinkedIn_KSrinivasan.txt 6.Definition TextGraph of connections (.txt) extracts essence of the aura surrounding a social profile. 7.Dense subgraph (core) of this connections textgraph is a clustering coefficient measure i.e how well the neighbours of a vertex have semantic connections among themselves. 8.Though social profiles are prone to Sybil links for artificially bumping up popularity, previous analysis is based on the assumption that humans do not link to another human unless they are trusted/befriended. 9.logs for this have been committed to: SocialNetworkAnalysis_PeopleAnalytics.log.20November2018 (degree experiential merit) SocialNetworkAnalysis_PeopleAnalytics.log2.20November2018 (recursive lambda function growth merit) 10.Recursive Lambda Function Growth merit logs have been truncated because of following recursion depth error in bintrees AVL Tree: RuntimeError: maximum recursion depth exceeded but the logs show few erratic cycles having "missiles" which are inferred from random walk paths printed in the logs. Also Names of the connections have not been filtered which misleads the WordNet. Previous error impedes full analysis based on core number. But there are some relevant cycles related to "head hunter","Ph.D","dissertation","entrepreneur","engineer","managers","information_technology" etc., giving a glimpse of the social circle's nature.
Srinivasan Kannan
More... almost 8 years ago
Theory - Complementary Sets (Beatty Sequencesi, Exact Cover) version of ABC Conjecture
Srinivasan Kannan
More... almost 8 years ago
Theory - Complement Diophantines and Beatty Theorem
Srinivasan Kannan
More... almost 8 years ago
Feature - Software Analytics - Cyclomatic Complexity
Srinivasan Kannan
More... almost 8 years ago
---------------------------------------------------------------------------------------------------------------- 616. (FEATURE) Software Analytics - Cyclomatic Complexity - Rewrite for merging two clauses - 14 November 2018 ---------------------------------------------------------------------------------------------------------------- 1. CyclomaticComplexitySparkMapReducer.py has been changed to merge True and False clauses for GraphFrames 2. Cyclomatic Complexity (Zeroth Betti Number) has been computed from GraphFrames stronglyConnectedComponents() as per definition of Cyclomatic Number in Topological Graph Theory 3. Cyclomatic Complexity (First Betti Number) has been computed as E-V+[Zeroth Betti Number] 4. New callgrind and kcachegrind DOT files for execution of command "ls" have been committed and included for Cyclomatic Complexity and Call graph GSpan mining: callgrind.ls.out.3104 kcachegrind_callgraph_ls.dot 5. logs for this have been committed to CyclomaticComplexitySparkMapReducer.log.GraphXFrames.14November2018
Srinivasan Kannan
More... almost 8 years ago
Theory - Diophantine Complexity Class D and RH
Srinivasan Kannan
More... almost 8 years ago
Theory - Pattern Mining Stream of LSH Partitions and Hash Table Chains - Correction
Srinivasan Kannan
More... almost 8 years ago
Theory - Pattern Mining Stream of LSH Partitions and Hash Table Chains
Srinivasan Kannan
More... almost 8 years ago
Theory - Complement Diophantine representation of Rationals - Correction
Srinivasan Kannan
More... almost 8 years ago
Theory - Diophantine Representation of Riemann Zeta Function - Reference
Srinivasan Kannan
More... almost 8 years ago
Feature - People Analytics - Unicode Errors and Another Example
Srinivasan Kannan
More... almost 8 years ago
-------------------------------------------------------------------------------------------------------------- Social Network Analysis - People Analytics - Another example and Unicode corrections -------------------------------------------------------------------------------------------------------------- 1. RecursiveGlossOverlap_Classifier.py has been changed to ignore unicode errors in text strings 2. New example Curriculum Vitae TeX (self) has been analyzed by Recursive Lambda Function Growth Algorithm for various intrinsic merit measure it prints. 3. Core classifier shows high relevance to human judgement and again Closeness and Betweenness centralities are better than Degree and PageRank centralities (classes like "Silver" are probably produced because of relevance of "Silver" to "Hyderabad") 4. RecursiveLambdaFunctionGrowth.dot file has been recreated. 5. logs for this have been committed to testlogs/SocialNetworkAnalysis_PeopleAnalytics.log.12November2018 6. Textual analysis of a CV is quite open-ended and unstructured - Nonetheless the classification by TextGraph core numbers gives reasonable estimate of crucial aspects in the resume
Srinivasan Kannan
More... almost 8 years ago
Updated BestPractices.txt
Srinivasan Kannan
More... almost 8 years ago
Updated Requirements.txt
Srinivasan Kannan
More... almost 8 years ago
Feature - People Analytics and Least Energy Intrinsic Merit of Social Profile
Srinivasan Kannan
More... almost 8 years ago
---------------------------------------------------------------------------------------------------------- 613. (FEATURE) People Analytics - HR Analytics Implementation and Least Energy Intrinsic Merit of a social profile - 9 November 2018 - related to 443, 572 ---------------------------------------------------------------------------------------------------------- (*) People Analytics is the generalized version of HR Analytics for analyzing any social profile vertex. (*) This commit adds SocialNetworkAnalysis_PeopleAnalytics.py python implementation to repository which at present analyzes LinkedIn profiles and computes: (*) Graph Tensor Neuron Network Intrinsic Merit of the profile text with no field parsing - this treats the social profile as an en masse text with no delineations and applies Recursive Lambda Function Growth algorithm to create lambda function composition tree of the resume text. Core classification and Centrality measures throw some interesting insights: (*) Core number 4 identifies the class of the social profile (*) Maximum merit random walk of the lambda composition tree approximately guesses the purport of the text in the social profile (*) Betweenness and Closeness Centralities estimate the top-ranked classes accurately compared to Degree and PageRank Centralities (*) Log Normal Fitness of the social profile after parsing the total work experience and academic stints of the profile - for linkedin this inverse log normal least energy fitness is 1/(logE + logV) with no Wealth estimates - Wealth is caused by E(Education) and V(Work/Valour) as far as professional profiles are concerned. Lowest Value of Least Energy Fitness implies Highest Fitness. Parsing is done by parse_profile() function which implements a switch design pattern - stint polymorphs from work to academic tenures. (*) Log Normal Experiential Intrinsic Fitness of the social profile (by applying mistake bound learning differential equation for recursive mistake correction tree described earlier) - log normalized because of huge exponents (*) Because of limitations due to Updated LinkedIn Scraping Policy (https://www.forbes.com/sites/forbestechcouncil/2017/09/20/linkedin-vs-hiq-ruling-casts-a-long-shadow-over-the-tech-industry/#19f535f35e6c), this is only implemented as a non-crawled example for LinkedIn profile of self: https://www.linkedin.com/comm/in/srinivasan-kannan-608a671 (pdf and text version of pdf by pdf2txt) (*) datetime python package is used to compute timedelta(s) and a regex matcher isdaterange() function has been implemented for parsing date ranges in the profile: (*) Work Experience Date Ranges have the Regex: <Month> <Year> - <Month> <Year> (*) Academic Date Ranges have the Regex: <Year> - <Year> and months are implicit (*) This implementation does not assume a Web Crawler and Social Profile Datasource and only requires a file input (datasource, file type and file name are passed in as arguments to parse_profile()) - optionally, crawled social profiles can be invoked as file arguments to parse_profile(). Presently only linkedin datasource, text file types are supported. PDF parsing has been implemented by PyPDF2 but there are IndirectObject related errors.
Srinivasan Kannan
More... almost 8 years ago
Theory - Many Unit Heaviside Step Function - Randomized Rounding
Srinivasan Kannan
More... about 8 years ago
Planar Point Location - Hyperbolic Tile Segments as adjacent Arrangements
Srinivasan Kannan
More... about 8 years ago