Ki2dZddlmZddlZddlmZmZddlmZm Z eje Z d d dZ d d dZdd Z d d d Z dd Zy)zcGraph analysis: hub detection, bridge nodes, knowledge gaps, surprise scoring, suggested questions.) annotationsN)Counter defaultdict) GraphStore_sanitize_namec f|j}t}t}|D]0}||jxxdz cc<||jxxdz cc<2|j d}|j }g}|D]} | j } |j| d} |j| d} | | z} | dk(r>|jt| j| j | j| j| | | |j| d|jdd|d|S) zFind the most connected nodes (highest in+out degree), excluding File nodes. Returns list of dicts with: name, qualified_name, kind, file, in_degree, out_degree, total_degree, community_id rT exclude_filesr)namequalified_namekindfile in_degree out_degree total_degree community_idc |dS)Nrxs f/home/jay/workspace/scripts/.codegraph-venv/lib/python3.12/site-packages/code_review_graph/analysis.pyz find_hub_nodes..3s a/keyreverseN) get_all_edgesrsource_qualifiedtarget_qualified get_all_nodesget_all_community_idsr getappendrr r file_pathsort)storetop_nedgesrrenodes community_mapscorednqnindoutdtotals rfind_hub_nodesr3s:    !E ' I!(J +1%%&!+&!$$%*%+   d  3E//1M F     mmB"~~b!$d  A:  "166*..FFKK!)--b1   $ KK/K> &5>rc ddl}|j}|j}|dkDr!td|}|j ||d}n|dkDr|j |d}ngS|j }|j dDcic]}|j|} }g} |jD]\} } | dks| | vr| | }|jd k(r%| jt|j|j|j|jt| d |j| d | j!d d | d|Scc}w)a1Find nodes with highest betweenness centrality. These are architectural chokepoints that sit on shortest paths between many node pairs. If they break, multiple communities lose connectivity. Returns list of dicts with: name, qualified_name, kind, file, betweenness, community_id rNiiT)k normalized)r6r File)r r rr betweennessrc |dS)Nr9rrs rrz#find_bridge_nodes..is q/rr)networkx_build_networkx_graphnumber_of_nodesminbetweenness_centralityr"r!r itemsrr$rr r%roundr#r&) r'r(nxnxgn_nodesr5bcr,r.node_mapresultsr/scores rfind_bridge_nodesrI7sp  % % 'C!!#G~ W   & &saD & A 1  & &st & < //1M$$4$8  !H GXXZ  E A:8+  RL 66V  "166*..FFKK ?)--b1     LL/L> 6E?-s Ec 0|j}|jd}|j}t}t }|D][}||j xxdz cc<||j xxdz cc<|jdk(sA|j|j ]g}|D]l}|j|jd} | dks%|jt|j|j|j|j| dn|j} g} | D]P} | jdddks| j| d | jd d | jddd Rg} |D]}|j|jd} | d k\s%|j|vs4|j rA| jt|j|j|j|j| d| j#ddt%t}|D]>}|j|j}|!||j|j@g}| D]} | d }|j|t }t'|dk(s1| jdddk\sG|j|| jd d | jddt)t+|d|dd| | dd|dS)aDIdentify structural weaknesses in the codebase graph. Returns dict with categories: - isolated_nodes: degree <= 1, disconnected from graph - thin_communities: fewer than 3 members - untested_hotspots: high-degree nodes with no TESTED_BY edges - single_file_communities: entire community in one file Tr r TESTED_BYr)r r rrdegreesizeidr )rr rMc |dS)NrLrrs rrz%find_knowledge_gaps..s akrrN)rr rMr2)isolated_nodesthin_communitiesuntested_hotspotssingle_file_communities)rr!r"rsetrr raddr#r r$rr r%get_communities_listis_testr&rlennextiter)r'r)r+r,rL tested_nodesr*isolatedr.d communitiesthincrW comm_filescid single_filefiless rfind_knowledge_gapsrjms     !E   d  3E//1M%YF UL 1q!!"a'"q!!"a'" 66[   Q// 0 1H   JJq'' + 6 OO&qvv."#"2"2    ,,.K D  55 a  KK !$fb)fa(    JJq'' + F$$L8  $ $&qvv."#"2"2  &   !4 '2#&6J - 0 01 ? sO   ,- K  gsCE* u:?quuVQ/14    #fb)fa(T%[)    #3B- .s3#.  rc |j}|jd}|j}|Dcic]}|j|}}t }|D]0}||j xxdz cc<||j xxdz cc<2|jD cgc] } | dkDs |  } } | sgSt| t| dz} t| dzd} g} |D]w}|j|j }|j|j }|r|s?|jdk(s|jdk(r^d }g}|j|j }|j|j }||||k7r|d z }|jd d |jvr|jjd ddnd}d |jvr|jjd ddnd}|r|r||k7r|dz }|jd|j|j d}|j|j d}|dkr|| k\s |dkr|| k\r|dz }|jd|j |j k7r%|jdk(r|dz }|jd|jdk(r%|jdk(r|dz }|jd|dkDs| jt#|j$|j t#|j$|j |jt'|d|||d z| j)dd| d |Scc}wcc} w)a~Find edges with high surprise scores. Detects unexpected architectural coupling based on: - Cross-community: source and target in different communities - Cross-language: different file languages - Peripheral-to-hub: low-degree node to high-degree node - Cross-file-type: test calling production or vice versa - Non-standard edge kind for the node types Tr rrrN r7gNg333333?cross-community.rPg?zcross-languagezperipheral-to-hubCALLSg333333?zcross-test-boundaryTypezunusual-edge-kind) sourcertargetr edge_kindsurprise_scorereasonssource_communitytarget_communityc |dS)Nrvrrs rrz-find_surprising_connections..*s a()rr)rr!r"r rrr valuessortedr]maxr#rr$r%rsplitr\rr rAr&)r'r(r)r+r,r.rFrLr*rbdegrees median_deghigh_deg_threshold scored_edgessrctgtrHrwsrc_cidtgt_cidsrc_langtgt_langsrc_degtgt_degs rfind_surprising_connectionsrs    !E   d  3E//1M-23  !#3H3%YF (q!!"a'"q!!"a'"( !--/3QQUq3G3  W!23JZ!^R0L ?ll1--.ll1--.#  88v V!3  ##A$6$67##A$6$67  'w& SLE NN, - cmm# MM a ( ,)+  cmm# MM a ( ,)+  X%9 SLE NN+ ,**Q//3**Q//3 \g);;qL#55 SLE NN. / ;;#++ %!&&G*; TME NN0 1 66W V!3 TME NN. / 19   (2$%$6$6(2$%$6$6VV"'q/"$+$+ ! k?B )4  i44sM4 M9)M9c g}t|d}|D]"}|jdd|dd|ddd $t|d}|j}|Dchc]}|jd k(r |j }}|D]0}|d|vs |jd d |dd |dd|ddd 2t |d} | D]<} d| dvs |jdd| dd| dd| dd| dd | ddd >t|} | dddD]+} |jdd | dd!| d"d#d$| d%d&d -| d'ddD](}|jd(d|dd |d)d*|ddd *|Scc}w)+amAuto-generate review questions from graph analysis. Categories: - bridge_node: Why does X connect communities A and B? - isolated_node: Is X dead code or dynamically invoked? - low_cohesion: Should community X be split? - hub_risk: Does hub node X have adequate test coverage? - surprising: Why does A call B across community boundary? rN)r( bridge_node'r z`' is a critical connector between multiple code regions. Is it adequately tested and documented?r high)categoryquestionrtpriorityrKhub_riskz Hub node 'z' has rz> connections but no direct test coverage. Should it be tested?rnrwsurprising_connectionrsz ' (community rxz ) calls 'rtryz ). Is this coupling intentional?rmediumrVNrlthin_communityz Community 'z ' has only rMz0 member(s). Should it be merged with a neighbor?z community:rlowrWuntested_hotspotrLz2 connections but no test coverage. Is this a risk?)rIr$r3rrrrrj) r' questionsbridgesbhubsr)r*testedh surprisessgapsres rgenerate_suggested_questionsr/s~I Q/G   %AfI;45()     %q )D    !E !&&K/ F   f ,   & 6 6()* ,-"   ,E;I   ) ,   3( }M+,-.( }M+,-.,- ./$    u %D $ %bq )  (ai[ V9+*+#1^#4"56    % &r *  *AfI;fQx[M2"#()      }s"E5)rm)r'rr(intreturn list[dict])r'rrzdict[str, list[dict]]))r'rrr)__doc__ __future__rlogging collectionsrrgraphrr getLogger__name__loggerr3rIrjrrrrrrs*#,-   8 $&T%'3 3!33lXx%'d d !d d N] ]]r